2016_UNESCO-IHE_PHD_THESIS_BRUINS_i

Transcription

Manganese Removal
from Groundwater
Role of Biological and Physico-Chemical
Autocatalytic Processes
Jantinus Bruins
MANGANESE REMOVAL FROM GROUNDWATER:
ROLE OF BIOLOGICAL AND PHYSICO-CHEMICAL
AUTOCATALYTIC PROCESSES
Cover image:
Impressions of the initially biologically formed Birnessite (green) by Pseudomonas sp. (yellow), and its subsequent
autocatalytic physico-chemical form (blue) during ripening of manganese-removal filters (made by Yness M. Slokar,
UNESCO-IHE).
Original Scanning Electron Microscopy (SEM) photos of biological and physico-chemical Birnessite: Arie
Zwijnenburg (Wetsus).
Original SEM photos of Pseudomonas sp.: Jelmer Dijkstra (Wetsus).
Cultivation and biological sample preparation of Pseudomonas sp. for SEM analysis: Pim Willemse (WLN).
ROLE OF BIOLOGICAL AND PHYSICO-CHEMICAL
AUTOCATALYTIC PROCESSES
DISSERTATION
Submitted in fulfillment of the requirements of
the Board for Doctorates of Delft University of Technology
and
of the Academic Board of the UNESCO-IHE
Institute for Water Education
for
the Degree of DOCTOR
to be defended in public on
Tuesday, 28th June 2016, at 15:00 hours
in Delft, the Netherlands
by
Jantinus Henderikus BRUINS
Master of Environmental Science, Open University Heerlen
born in Eext, the Netherlands
iii
This dissertation has been approved by the
promotor
: Prof. dr. ir. M.D. Kennedy and
copromotor : Dr. ir. B. Petrusevski
Composition of Doctoral Committee:
Chairman
Vice-Chairman
Prof. dr. ir. M.D. Kennedy
Dr. B. Petrusevski
Rector Magnificus TU Delft
Rector UNESCO-IHE
UNESCO-IHE / TU Delft, promotor
UNESCO-IHE, copromotor
Independent members:
Prof. dr. ir. W.G.J. van der Meer
Prof. dr. ir. J.P. van der Hoek
Prof. dr. V. L. Snoeyink
Dr. P.S. Hofs
Delft University of Technology
Delft University of Technology
University of Illinois, USA
Evides Water, the Netherlands
Prof. dr. ir. M.E. McClain
TU Delft / UNESCO-IHE, reserve member
CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business
© 2016, Jantinus Bruins
Although all care is taken to ensure integrity and the quality of this publication and the information herein, no
responsibility is assumed by the publishers, the author nor UNESCO-IHE for any damage to the property or
persons as a result of operation or use of this publication and/or the information contained herein.
A pdf version of this work will be made available as Open Access via http://repository.tudelft.nl/ihe This version
is licensed under the Creative Commons Attribution-Non Commercial 4.0 International License,
http://creativecommons.org/licenses/by-nc/4.0/
Published by:
CRC Press/Balkema
PO Box 11320, 2301 EH Leiden, the Netherlands
[email protected]
www.crcpress.com – www.taylorandfrancis.com
ISBN 978-1-138-03002-2
iv
Acknowledgements/dankbetuiging
Geheel tegen de schijnbare gewoonte in bij het verwoorden van een dankbetuiging, wil ik niet eindigen,
maar beginnen met diegenen, die vóór en tijdens mijn studie een zeer belangrijke rol hebben gespeeld. Het
noemen van de partner zal voor velen een cliché zijn, maar ik kan u verzekeren hoe anders dit is als het je
zelf betreft. Want het was zeker niet alleen tijdens deze studie dat Martha mijn steun en toeverlaat is geweest
en nog altijd is. Al sinds onze eerste kennismaking in 1984 heeft ze me niet anders meegemaakt dan
studerende. Ook toen Marcel en Ronald onderdeel uit gingen maken van ons gezin, was zij het die verreweg
het grootste deel van de opvoeding en zorg voor onze zonen op zich heeft genomen. Als zeer toegewijde
moeder, heeft zij zich hierdoor volledig weggecijferd en daarmee haar werkzame carrière opgeofferd. Dit
alles om mij de kans te geven me te ontplooien en te komen waar ik nu sta. Martha, dit verdient alleen maar
mijn diepste respect, waar ik jou - uit de grond van mijn hart - enorm dankbaar voor ben.
Ik noemde al Marcel en Ronald. Mijn “leven lang leren” heeft er ongetwijfeld voor gezorgd dat ik er niet
altijd voor jullie was, terwijl dit wel zo had moeten zijn, zoals jullie dat ook van een goede vader hadden
mogen verwachten. Dit spijt me, maar ik troost me met de gedachte dat dit gemis, ruimschoots is
gecompenseerd door jullie fantastische moeder.
Nog meer aan de basis van mijn carrière als “eeuwige” student, stonden mijn ouders. Zij hebben mij altijd
gestimuleerd, waar en waarmee ze ook maar konden om een opleiding te volgen. Dit begon al op de
middelbare school en werd met evenveel enthousiasme doorgezet tijdens mijn promotietraject. Pap, helaas
heb je de voltooiing van dit traject niet mee mogen maken, maar ik ben heel blij dat de start je nog wel
gegeven is. Mam, je weet niet hoe ongelofelijk blij ik ben dat jij dit slotstuk wel mee kunt maken. Dank voor
alles wat jullie voor mij gedaan hebben en ik hoop dat je een beetje trots kunt zijn. Het is dezelfde trots, die
ik ook bespeurde bij mijn schoonouders. Helaas heeft ook mijn schoonvader het einde van mijn
promotietraject net niet mogen meemaken.
Natuurlijk was het uitvoeren en afronden van mijn promotietraject een stuk moeilijker, zo niet onmogelijk
geweest als mijn werkgever WLN (verpersoonlijkt door Hilde en Gerrit) en hiermee ook de twee
moederbedrijven; Waterleidingmaatschappij Drenthe en Waterbedrijf Groningen mij niet de ruimte en de
financiële steun hadden gegeven die ik kreeg. Ik hoop dat dankzij dit onderzoek, voor hen en met hen de
hele drinkwaterwatersector, de klassieke grondwaterzuivering en in het bijzonder de opstart van het
ontmanganingsproces in de toekomst op een snellere en nog duurzamere wijze kan worden uitgevoerd.
Mijn promotieonderzoek werd uiteindelijk uitgevoerd bij UNESCO-IHE in Delft. Branislav, ik ben je
dankbaar voor je inzet, stimulans en immer positief kritische houding als het ging over het werk dat ik
uitvoerde, de “abstracts” en “papers” die ik moest indienen. En natuurlijk vergeet ik zeker niet de
persoonlijke rondleiding door Belgrado, die ik samen met Salifu en Valentine van jou en je echtgenote heb
gehad rond een congres. Dit blijft een mooie en bijzonder waardevolle herinnering. En in de loop van de
jaren kregen onze gesprekken ook vaak een persoonlijke(re) noot. Yness, thank you very much for
everything you have done for me. Your experience in writing a good paper for a peer reviewed journal was
very valuable for me, and I know that you had to stay calm and patience over and over again, to learn me a
few of your skills. Your creativity was also very helpfull. Tijdens het schrijven van artikelen, maar ook voor
dit proefschrift heb ik de nodige adviezen van Joop gehad. Jouw jarenlange ervaring heb ik direct kunnen
omzetten in een enorme kwaliteitsimpuls wat betreft de manuscripten. Tot slot heeft Maria als promotor
een belangrijke stempel gedrukt op dit onderzoek door richting te geven, maar vooral ook door grenzen te
stellen. Als ze dit laatste niet had gedaan, was ik waarschijnlijk nog steeds bezig geweest met praktisch
onderzoek.
v
Het gehele promotietraject is dan wel uitgevoerd “onder de vlag” van UNESCO-IHE, maar het grootste en
ook het belangrijkste deel van het praktische onderzoek is uitgevoerd bij de collega’s van het Vlaamse
waterbedrijf Pidpa. De directie en medewerkers van Pidpa ben ik dan ook zeer erkentelijk voor hun steun
en toewijding tijdens mijn onderzoek. Hoewel binnen Pidpa velen een bijdrage hebben geleverd wil ik toch
een paar collega’s in het bijzonder noemen. Hun bijdrage is cruciaal geweest in dit geheel. Het balletje is
natuurlijk gaan rollen door de goede contacten die ik, sinds het laatste decennium van de vorige eeuw, heb
met Koen. Op het moment dat mijn promotieonderzoek en het onderwerp ervan ter sprake kwam, was
Koen meteen enthousiast en had wel een locatie beschikbaar voor onderzoek. Dit aanbod heb ik uiteraard
in dank aanvaard en met hulp van mijn voormalige collega Chris, mijn huidige collega Karl (toen Vitens),
“de andere” Koen van Pidpa en een aantal medewerkers van de locatie Grobbendonk, werd de pilot
opgebouwd. Bij de begeleiding/advisering van dit pilotonderzoek heb ik enorm veel steun gehad van beide
Koenen. Voor een groot deel werd dit pilotonderzoek ook door hen uitgevoerd, uiteraard met hulp van
Martine en Ann.
Het waren overigens niet alleen mijn collega’s Chris en Karl van WLN die een bijdrage aan mijn
promotietraject hebben geleverd, maar met hen vele anderen. Natuurlijk mijn directe collega’s van
technologie, die door mijn afwezigheid, misschien af en toe ”een stap harder moesten lopen”, maar die
daarentegen ook vele dagen niet mijn “gezeur en geklaag” hoefden aan te horen. Gerhard (ook als mede
auteur), Marsha, Pim en Arnout hebben veel werk verzet tijdens het deel waar de moleculaire microbiologie
om de hoek kwam kijken. Verschillende laboratorium collega’s van de diverse afdelingen heb ik mogelijk
schrik aangejaagd toen ik zelf laboratoriumanalyses kwam uitvoeren. En natuurlijk bedank ik ook mijn
collega’s, die tijdens het uitvoeren van hun werkzaamheden vaak niet in de gaten hadden dat ze een bijdrage
leverden aan mijn onderzoek. Veel van de te onderzoeken monsters zijn aangeleverd door Arjan (en zoals
je ziet Arjan, mijn dankwoord richting jou heeft een vergelijkbaar lettertype gekregen als het dankwoord aan
een ieder........dat valt niet tegen toch?).
Voor een aantal specialistische onderzoeken ben ik geholpen door verschillende deskundigen. Het
onderzoek met de Electron Paramagnetic Resonance (EPR) heb ik zelf mogen uitvoeren bij de Rijks
Universiteit Groningen, na een “spoedcursus EPR” door Dr. Wesley Browne. Onderzoeken met XRD
werden uitgevoerd door Ruud Hendrickx van de TU Delft. Ton heeft bij Wetsus heel veel onderzoek gedaan
met Raman spectroscopie en Jelmer heeft tegen het eind van mijn promotieonderzoek een aantal SEMopnamen gemaakt. En hoewel deze opnamen prima waren, had ik toch heel graag gehad dat ook deze laatste
SEM opnamen, net als alle andere, gemaakt waren door Arie. Het heeft helaas niet zo mogen zijn. Arie, ik
koester de fijne uren samen achter de SEM. De gesprekken over het werk, de vakanties, de foto’s en onze
families. Maar we begonnen natuurlijk met koffie, want zoals je altijd zei: “dat apparaat moet toch eerst even
opwarmen”. Jouw werk zal herkenbaar blijven in het onderzoek van velen en op deze wijze hoop ik de
herinnering aan jou ook voor altijd te laten voortleven. Arie, bedankt voor wie je was en wat je voor mij
hebt gedaan!
Vanuit het “werkveld” werd mijn promotietraject van de nodige ondersteuning voorzien, door een
begeleidingsgroep van experts. Hiervan maakten Koen en Joop ook deel uit en verder bestond de groep uit:
Ans (RIVM), Bas (KWR), Ben (De Watergroep) en Jacques (Vitens). Een ieder heeft hier op zijn of haar
wijze een waardevolle en belangrijke bijdrage geleverd aan het geheel. Met Jacques reisde ik vaak samen naar
Delft en tijdens de ‘lange’ reis hadden we vaak boeiende en interessante gesprekken, waardoor de afstand
Zwolle - Delft soms juist wel heel erg kort leek. Van KWR hebben ook Dirk en Bart (beiden als mede
auteurs) een waardevolle bijdrage geleverd. Ook verschillende collega’s van Nederlandse, Vlaamse en Duitse
waterbedrijven hebben informatie verstrekt die belangrijk was voor de inventarisatie van “ontmanganing in
de praktijk”. Dit heeft uiteindelijk heeft geleid tot de 1e publicatie.
vi
Behalve dat ik zelf de nodige onderzoeken heb uitgevoerd, zijn er ook een aantal MSc studenten geweest
die een deel voor hun rekening hebben genomen. The first student was Edilberto Ayala Baquero from
Colombia and student at Wetsus. Chris was de 2e student en hij volgde een opleiding aan Van HallLarenstein. Next students were all from UNESCO-IHE: Ahmed Abdullah Saif Al-Abri and Younis
Sulaiman Hamed AlZakwani, both from Oman, Nicholaus Angumbwike Njumbo from Tanzania and
Clement Ndungutse from Rwanda. De student die het geheel afsloot was Nicolas Soenens van de
Universiteit van Antwerpen.
Tot slot wil ik een ieder bedanken, die ik niet heb genoemd in dit dankwoord, maar die toch een belangrijke
bijdrage heeft geleverd aan het welslagen van mijn promotieonderzoek.
BEDANKT
THANK YOU
Jantinus Bruins
Juni, 2016
Assen, Nederland
vii
Voor de twee belangrijkste vrouwen in mijn leven.
For the two most important women in my life
Summary
Groundwater is the predominant source of drinking water globally. However, untreated groundwater
contains compounds that are undesirable in drinking water, such as methane, ammonia, iron and manganese.
The presence of manganese in drinking water is detrimental, because of health, aesthetic and practical
reasons. Increased levels of manganese in drinking water are associated with neurologic symptoms,
especially in young children. Moreover, a synergistic toxic effect exists in combination with arsenic.
In European countries, health related effects caused by manganese in drinking water are negligible and the
main problems are of aesthetic and practical concern, particularly in groundwater treatment.
In The Netherlands, manganese is removed by conventional groundwater treatment consisting of aeration
and rapid (sand) filtration. Such a treatment process is easy to operate, cost effective and sustainable as it
does not require the use of strong oxidants such as O3, Cl2, ClO2 and KMnO4 with the associated risk of
by-product formation and over or under dosing. However, the application of aeration-filtration has also
some drawbacks. In particular, the long ripening time required for filter media is a major concern. Currently,
water supply companies have to waste large volumes of treated water, which reduces the sustainability of
the process. In addition, the costs associated with filter media ripening (man power, electricity, operational
and analyses costs) are high. Consequently, this thesis describes the removal of manganese from
groundwater by microbial and physico-chemical auto-catalytic processes, with a specific focus on revealing
the mechanisms involved in filter media ripening with the aim to shorten or even completely eliminate filter
media ripening.
Initially, data was collected from over 100 full-scale groundwater treatment plants, mainly in The
Netherlands, Belgium and Germany and an inventory of the most important (water quality and operational)
parameters required for successful manganese removal by aeration-rapid sand filtration, was made.
Univariate statistics and assessment of available data (from over 100 plants) indicated that a very effective
manganese removal efficiency in the first aeration-filtration stage with simultaneous removal of iron and
ammonia, could be achieved under the following conditions:
•
•
•
•
•
•
NH4+ removal efficiency
iron loading per filter run
pH of filtrate
filtration rate
empty bed contact time
oxygen in filtrate
:
:
:
:
:
:
> 85%
< 2.7 kg Fe/m2
> 7.1
< 10.5 m/h
> 11.5 min
≥ 1 mg/L
An important step in revealing the processes and mechanisms involved in ripening of the filter media was
the identification of the manganese oxide present. To characterize and identify the manganese oxide present
on filter media coatings, the following techniques were used: Röntgen diffraction (XRD), Raman
spectroscopy, Scanning Electron Microscopy and Energy Dispersive X-ray analysis (SEM-EDX) and
Electron paramagnetic resonance (EPR). With these techniques, an amorphous type of manganese oxide:
Birnessite was identified in all samples examined. Birnessite has excellent properties to adsorb ions, such as
Mn2+, and has highly auto-catalytic oxidative properties. Both of these properties make Birnessite extremely
suitable to remove Mn2+ from groundwater in an effective and efficient way.
xi
The next part of the research focused on investigating the process of filter media ripening. Therefore, pilot
tests were conducted at the water treatment plant Grobbendonk (Pidpa, Belgium). During the ripening
process, samples of the filter media as well as the backwash water were collected and analyzed. Applying
the aforementioned techniques, the formed manganese oxide was characterized and identified. Moreover,
with the combination of SEM and EPR it was possible to distinguish whether the Birnessite was formed
biologically or physico-chemically. These techniques showed that at the beginning of the ripening process
the produced Birnessite was of biological origin. As filter ripening progressed and a coating developed on the
filter media, the produced Birnessite became predominantly of physico-chemical origin. After approximately
500 days, all Birnessite present on the filter media was produced physico-chemically.
Throughout the whole ripening period manganese oxide particles collected from filter backwash water were
consistently of biological origin, suggesting that biological oxidation of adsorbed manganese took place
throughout the filter run. Consequently, the bacteria population present in (freshly ripened) manganese
removal filters was examined. For this purpose, the following molecular DNA analyses were used: “next
generation DNA sequencing”, qPCR and MALDI-TOF MS analysis.
The “next generation DNA sequencing” analyses, showed a bacteria population shift during the start up
phase of the manganese removal process. In the filtrate of the iron removal filter as well as in the feed of
the manganese removal filter, the iron oxidizing genus Gallionella was dominant (> 97% of the total bacteria
population), whereas the backwash water of the manganese removal filter comprised only 12.4% of the
genus Gallionella. Bacteria of the genus Nitrospira and the genus Pseudomonas, 25.7 % and 14.3%, respectively,
were also present in the manganese removal filter. However, 47.6% of the bacteria population in the
manganese oxidizing column, consisted of small groups of bacteria which remained unknown.
Nitrospira is known to be involved in the oxidation of nitrite to nitrate and therefore its presence is expected,
because in addition to manganese, ammonia was also oxidized in this filter. Pseudomonas sp. and in particularly
P. putida is known to be capable of oxidizing Mn2+.
However, qPCR established that the presence of Pseudomonas putida was very limited. Less than 0.01% of the
genus Pseudomonas present was of the species Pseudomonas putida. After successive culturing, some strongly
related Pseudomonas species (amongst others: P. gessardii, P. grimontii and P. koreensis) were identified with
MALDI-TOF analysis. At GWTP Grobbendonk, Pseudomonas sp. is most likely the manganese oxidizing
bacterium genus playing an important role in the start up phase of filter media ripening. However, it is not
known whether this bacterium genus is operating alone or as part of a microbial consortium.
A study was carried out to assess the potential of manganese oxide-coated filter media (MOCS/MOCA) to
reduce the ripening time of filters with virgin media. The addition of a layer of fresh MOCA to the filter
eliminated the ripening time completely, while a layer of dry MOCS introduced to a virgin sand filter did
not significantly affect the ripening period. Both tests were applied at different locations, with different
operational and water quality parameters.
Based on the aforementioned findings, a pilot study with fresh and dry coated filter media was performed
under controlled conditions. Virgin filter media (sand and anthracite) were used as reference materials. From
this study it was concluded that, with comparable process conditions, the duration of filter media ripening
with virgin sand and anthracite was similar.
Furthermore, it was shown that freshly prepared manganese oxide coated filter media has excellent
properties to enhance the ripening process and was capable of eliminating ripening time completely. On the
xii
other hand, the use of dry MOCS, which adsorbs Mn2+ temporarily, did not have a significant impact on
the ripening time.
The effect of backwash frequency on filter media ripening was examined at pilot scale. This study confirmed
that more frequent filter backwashing negatively affected filter media ripening time with virgin media. Thus,
backwash frequency was shown to be a key factor in the start up of new filters for manganese removal. The
influence of backwashing became less pronounced as filter ripening progressed, due to the development of
a thicker layer of biomass and/or auto-catalytically active Birnessite on the media surface. Backwashing
showed very little impact on filter ripening time (manganese removal efficiency) when a layer of fresh
MOCA/MOCS was used. The backwash frequency depends on the amount of oxidized Fe2+ (Fe(OH)3)
which is retained by the filter. Consequently, the iron concentration in the feed water and the iron loading
are also key factors influencing the ripening time of manganese removal filters.
Prior to this research, it was believed that manganese removal by conventional aeration-filtration was a
physico-chemical process dominated by the formation of the manganese oxide Hausmannite and that
microbial processes were also involved. This thesis presents new information revealing the mechanisms and
processes involved in the start up of filters with virgin media in the removal of manganese from
groundwater. It was clearly shown that the predominant manganese oxide contributing to filter media
ripening was Birnessite. It was also found that Birnessite formed at the start of the ripening process was of
biological origin while as ripening progressed, Birnessite formed became predominantly of physico-chemical origin.
Based on the knowledge presented in this thesis, water supply companies can take measures to optimize the
filter ripening process, thereby reducing the ripening time. This can be achieved by creating conditions
favouring the growth of manganese oxidizing bacteria, for example by limiting the frequency of backwashing
(e.g., by limiting the iron loading of the filter).
Finally, filter media ripening can be completely eliminated by the addition of freshly prepared MOCS/MOCA,
containing Birnessite, to the filter.
xiii
xiv
Samenvatting
Grondwater is wereldwijd de belangrijkste bron voor de productie van drinkwater. In grondwater zijn van
nature verschillende bestanddelen aanwezig, die in drinkwater niet gewenst zijn, zoals; methaan, ammonium,
ijzer en mangaan.
De aanwezigheid van mangaan is ongewenst om gezondheidskundige, esthetische en praktische redenen.
Verhoogde concentraties mangaan kunnen vooral bij jonge kinderen aanleiding zijn voor neurologische
problemen. In combinatie met arseen kan er sprake zijn van een versterkt toxisch effect. In West-Europa
zijn de gezondheidskundige effecten ten gevolge van de aanwezigheid van mangaan in drinkwater
verwaarloosbaar en zijn de problemen vooral van esthetische en praktische aard. Deze praktische problemen
doen zich met name voor bij de zuivering van het grondwater. Zuivering van grondwater vindt in Nederland
vooral plaats door beluchting, gevolgd door filtratie. Dit is een eenvoudige, goedkope en duurzame vorm
van waterzuivering, omdat hierbij geen gebruik wordt gemaakt van chemicaliën zoals O3, Cl2, ClO2 en
KMnO4 ten behoeve van de oxidatie. Dit type zuivering kent dan ook niet de nadelen die verbonden zijn
aan het gebruik van deze sterke oxidatoren, zoals de vorming van bijproducten en over- of onder dosering.
Met betrekking tot de verwijdering van mangaan kent de toepassing van deze grondwaterzuiveringstechniek
ook een aantal nadelen, waarvan de lange rijpingstijd van nieuw filtermateriaal de belangrijkste is. De rijping
van nieuw filtermateriaal, voordat een volledige mangaanverwijdering is gerealiseerd, duurt over het
algemeen enkele maanden tot soms meer dan een jaar. Voor waterbedrijven leidt het rijpingsproces tot een
verlies van gezuiverd water, additionele kosten (arbeid, laboratorium, etc.) en een verlies aan
productiecapaciteit. Het is dan ook van belang het filterrijpingsproces zo kort mogelijk te houden.
Het onderwerp van het onderzoek beschreven in dit proefschrift is de verwijdering van mangaan uit
grondwater door middel van microbiologische en fysisch-chemische auto katalytische processen. De focus
ligt hierbij op het filterrijpingsproces, waarbij de aandacht in bijzonder was gericht op het ontrafelen van de
betrokken mechanismen en processen, met als doel de verkorting van de filterrijpingstijd.
Lange tijd is er van uitgegaan dat de verwijdering van mangaan uit grondwater, door middel van traditionele
beluchting-filtratie, een fysisch-chemisch proces was, met een belangrijke rol voor het mangaanoxide
Hausmannite. Later bleken ook microbiologische processen een rol te spelen. Ondanks het feit dat er veel
onderzoek is gedaan naar de mangaanverwijdering, zijn de processen en mechanismen die hierbij betrokken
zijn en hun onderlinge samenhang nog steeds niet volledig doorgrond. Doel van dit onderzoek was dan ook
het vergroten van de kennis en de wijze waarop het filterrijpingsproces tot stand komt. Door de ontrafeling
van het fenomeen filterrijping, kunnen oplosrichtingen om dit proces te verkorten geformuleerd worden.
Om het inzicht ten aanzien van mangaanverwijdering te vergroten, is bij de aanvang van dit onderzoek een
inventarisatie uitgevoerd bij meer dan 100 grondwaterzuiveringsinstallaties, met name in Nederland, België
en Duitsland. Op basis van deze inventarisatie was het mogelijk belangrijke parameters voor een succesvolle
ontmanganing vast te stellen. Door statistische correlaties is aangetoond dat volledige ontmanganing, in
aanwezigheid van ammonium en ijzer, mogelijk is wanneer voor een aantal parameters aan de volgende
criteria is voldaan:
• NH4+ - verwijderingsefficiëntie
:
> 85%
:
< 2,7 kg Fe/m2.FR
• Belading filter met ijzer, per filterrun
• pH van het filtraat
> 7,1
:
• filtratiesnelheid
:
< 10,5 m/h
:
> 11,5 min
• Schijnbare verblijftijd
• Zuurstofgehalte in filtraat
:
≥ 1 mg/l
xv
Een belangrijke stap in het ontrafelen van de processen en mechanismen die van belang zijn bij de start van
het filterrijpingsproces is het vaststellen van het type mangaanoxide dat hierbij betrokken is. Ten behoeve
van de karakterisatie en identificatie van mangaanoxide, aanwezig in filter media coatings, is gebruik gemaakt
van een aantal analysetechnieken, te weten:
•
•
•
•
Röntgen diffractie (XRD);
Raman spectroscopie;
Elektronenmicroscopie en “Energy Dispersive X-ray analysis” (SEM-EDX);
Elektron paramagnetische resonantie (EPR).
Op basis van de combinatie van bovenstaande analysetechnieken is vastgesteld dat in alle onderzochte
monsters van filter media coatings, een amorf mangaanoxide van het type Birnessite aanwezig was. Van
Birnessite is bekend dat het een mangaanoxide is met uitstekende adsorptie eigenschappen voor diverse
ionen, waaronder Mn2+. Verder heeft Birnessite ook auto katalytisch oxidatieve eigenschappen. Beide
eigenschappen maakt Birnessite uitermate geschikt voor een effectieve en efficiënte verwijdering van
mangaan uit grondwater.
Een volgende stap in het onderzoek was het volgen van het rijpingsproces. Hiervoor is een proefinstallatie
onderzoek uitgevoerd op productielocatie Grobbendonk van het Vlaamse waterbedrijf Pidpa. Gedurende
het gehele rijpingsproces zijn monsters genomen van zowel het filtermateriaal als het spoelwater. Met behulp
van de beschreven analysetechnieken is het gevormde mangaanoxide gekarakteriseerd en geïdentificeerd.
Bovendien is het met de combinatie van SEM en EPR vastgesteld of Birnessite, biologisch of fysischchemisch is gevormd. Op basis van deze technieken kon worden vastgesteld dat de vorming van het
mangaanoxide (Birnessite) op biologische wijze was gestart. Gedurende het rijpingsproces nam het aandeel
fysisch-chemisch gevormd Birnessite toe. Na ca. 450 - 550 dagen bleek de Birnessite, aanwezig op het filter
materiaal, volledig op fysisch-chemische wijze gevormd te zijn. Birnessite, aanwezig in het spoelwater, bleek
gedurende het gehele filterrijpingsproces echter vooral op biologische wijze te zijn geproduceerd.
Tijdens het pilotonderzoek is ook de bacteriepopulatie, aanwezig in de mangaanverwijderingsfilters,
onderzocht. Bij dit onderzoek is gebruik gemaakt van een aantal moleculaire analysetechnieken:
•
•
•
“Next generation DNA-sequencing”;
qPCR;
MALDI-TOF MS analyse.
Tijdens dit onderzoek is een duidelijke verschuiving van de bacteriepopulatie tijdens de start van het
ontmanganingsproces aangetoond. In het filter dat als voedingswater diende voor het
mangaanverwijderingsfilter waren met name bacteriën van het geslacht Gallionella aanwezig (>97% van de
totale bacterie populatie). Bacteriën, aanwezig in het spoelwater van het mangaanverwijderingsfilter, kort
nadat de ontmanganing volledig was, bestond nog voor “slechts” 12,4 % uit bacteriën van het geslacht
Gallionella. Verder waren de bacterie geslachten: Nitrospira (25,7%) en Pseudomonas (14,3%) aanwezig.
Ongeveer 47.6% van de bacteriepopulatie bestond uit kleine bacteriegroepen, die veelal niet volledig
gekarakteriseerd zijn. Van Nitrospira is bekend dat het in staat is nitriet om te zetten in nitraat als onderdeel
van de ammoniumoxidatie. De aanwezigheid van dit bacterie geslacht is dus verklaarbaar, omdat naast
mangaan ook ammonium werd omgezet in dit filter. Van Pseudomas sp., en in het bijzonder Pseudomonas
putida, is bekend dat het in staat is mangaan te oxideren.
xvi
Uit qPCR-analyses is echter gebleken dat de species P. putida, slechts in zeer geringe mate aanwezig was (<
0,01%). Met behulp van MALDI-TOF MS zijn een aantal nauw verwante Pseudomonas soorten aangetroffen,
onder meer: P. gessardii, P. grimontii en P. koreensis. Het lijkt er dan ook op dat het bacteriegeslacht Pseudomonas
betrokken is bij de start van het filterrijpingsproces. Tijdens dit onderzoek is het, onder gecontroleerde
laboratoriumcondities, niet gelukt geïsoleerde Pseudomonas soorten Mn2+ te laten oxideren, dit in tegenstelling
tot een laboratoriumstam van P. putida. Of mangaanoxidatie in de proefinstallatie van Grobbendonk een
solitaire actie is van deze bacteriegeslacht, dan wel een gezamenlijk proces met een consortium van ook
andere bacteriën, is niet vastgesteld tijdens dit onderzoek.
De effectiviteit van Birnessite in filter media met mangaanoxide coating (MOCS/MOCA) is vervolgens op
praktijkschaal getest. Verse MOCA was in staat de rijpingstijd tot nul te reduceren. Droog MOCS was niet
in staat de rijpingstijd te verkorten. Bovenstaande testen zijn echter uitgevoerd op twee verschillende locaties
onder verschillende operationele condities. Bovendien was ook de kwaliteit van het grondwater op beide
locaties verschillend.
Op basis van bovenstaande bevindingen zijn vervolgens, op proefinstallatieschaal, testen uitgevoerd met
droog en vers gecoat filtermateriaal onder dezelfde condities. Tijdens dit onderzoek is nieuw filter materiaal
(zand en antraciet) gebruikt als referentie. Uit deze experimenten is geconcludeerd dat er bij gebruik van
nieuw zand en nieuw antraciet, sprake was van vergelijkbare filterrijpingstijden. Verder is vastgesteld dat bij
toepassing van zowel vers MOCA als vers MOCS de filterrijpingstijd is gereduceerd tot nul. Droog gecoat
filtermedia leidde slechts tot een tijdelijke adsorptie van Mn2+.
Verder is tijdens dit onderzoek de invloed van de terugspoeling van een filter op het filterrijpingsproces
onderzocht. De frequentie van het terugspoelen van een filter had een belangrijke, negatieve, invloed op de
start van het filterrijpingsproces. Hiermee vervult de filterspoeling dus een sleutelrol en is daarmee een
zogenaamde sleutelfactor bij de start van het ontmanganingsproces.
Gedurende het filterrijpingsproces neemt de invloed van de filterspoelingen af, ten gevolge van de
ontwikkeling van een dikkere laag biomassa en/of auto katalytisch actief Birnessite.
Filterspoelingen hebben geen of slechts een marginale invloed op de mangaanverwijdering, wanneer gebruik
gemaakt wordt van een laag vers MOCS of MOCA.
De terugspoelfrequentie van een filter is vooral afhankelijk van de hoeveelheid geoxideerd Fe2+ (Fe(OH)3),
die wordt afgevangen in het filter (de filterbelading met ijzer). Ook de belading van het filter met ijzer en de
concentratie van Fe2+ in ruw water, kunnen hierdoor worden beschouwd als sleutelfactoren met betrekking
tot de start van het ontmanganingsproces.
Resumerend kan worden gesteld dat in dit proefschrift nieuwe en belangrijke inzichten zijn gepresenteerd
met betrekking tot het ontrafelen van processen en mechanismen die betrokken zijn bij (de start van) het
filterrijpingsproces.
Aangetoond is dat Birnessite, en niet Hausmanite, een cruciale rol speelt bij ontmanganing. De vorming van
Birnessite begint biologisch, maar gedurende het ontmanganingsproces blijkt de vorming van fysischchemisch auto katalytisch Birnessite dominant. Bij de biologische ontmanganing spelen naar alle
waarschijnlijkheid meerdere bacteriesoorten een rol, waaronder bacteriën van het geslacht Pseudomonas.
Op basis van deze inzichten is het voor waterbedrijven mogelijk, bij de engineering en in de dagelijkse
bedrijfsvoering, maatregelen te nemen om het filterrijpingsproces te optimaliseren. Deze maatregelen
kunnen zijn gericht op het creëren van optimale condities voor mangaanoxiderende bacteriën, waarbij met
xvii
name de frequentie van de filterspoeling wordt beperkt (b.v. door het beperken van de belasting van het
filter met ijzer). Verder kan gebruik gemaakt worden van vers gecoat filter media (MOCS/MOCA),
waardoor volledige eliminatie van de filterrijpingstijd realiseerbaar is.
xviii
Contents
Acknowledgements/dankbetuiging...........................................................................................................................V
Summary......................................................................................................................................................................XI
Samenvatting..............................................................................................................................................................XV
1
GENERAL INTRODUCTION ............................................................................................... 1
1.1
Manganese in groundwater and groundwater treatment in The Netherlands................................... 2
1.1.1
Manganese occurrence in groundwater .......................................................................................... 2
1.1.2
The relevance of manganese in drinking water and guideline values ........................................ 2
1.1.3
Groundwater treatment in the Netherlands .................................................................................. 3
1.2
Problem description regarding manganese removal in practice .......................................................... 4
1.3
Aim and research objectives of this thesis .............................................................................................. 6
1.4
Outline of this thesis................................................................................................................................... 7
1.5
References .................................................................................................................................................... 8
2 ASSESSMENT OF MANGANESE REMOVAL FROM OVER 100 GROUNDWATER
TREATMENT PLANTS............................................................................................................ 13
2.1
Abstract.......................................................................................................................................................14
2.2
Introduction ...............................................................................................................................................14
2.3
Materials and Methods .............................................................................................................................17
2.4
Results and Discussion.............................................................................................................................18
2.4.1
Multivariate statistics and univariate correlations .......................................................................18
2.4.2
NH4+ removal efficiency.................................................................................................................21
2.4.3
Effect of iron loading ......................................................................................................................22
2.4.4
Effect of (filtrate) pH ......................................................................................................................24
2.4.5
Other parameters .............................................................................................................................25
2.4.6
Effect of filtration rate (m/h) ........................................................................................................25
2.4.7
Effect of contact time and filter bed depth .................................................................................26
2.4.8
Effect of oxygen concentration .....................................................................................................26
2.4.9
Effect of filtration type (gravity or pressure)...............................................................................27
2.5
Conclusions ................................................................................................................................................28
2.6
Acknowledgements ...................................................................................................................................29
2.7
References ..................................................................................................................................................29
3 MANGANESE REMOVAL FROM GROUNDWATER: CHARACTERIZATION OF
FILTER MEDIA COATING ..................................................................................................... 33
3.1
Abstract.......................................................................................................................................................34
3.2
Introduction ...............................................................................................................................................34
3.3
3.3.1
Raman spectroscopy ........................................................................................................................38
3.3.2
XRD ...................................................................................................................................................38
3.3.3
SEM-EDX ........................................................................................................................................39
3.3.4
EPR ....................................................................................................................................................39
xix
3.4
3.4.1
Selection Raman spectroscopy settings ........................................................................................39
3.4.2
Characterization of MOCA and MOCS by Raman spectroscopy ...........................................41
3.4.3
XRD ...................................................................................................................................................43
3.4.4
SEM-EDX ........................................................................................................................................44
3.4.5
EPR ....................................................................................................................................................45
3.4.6
The importance of Birnessite formation on MOCA/MOCS for manganese removal ........47
3.5
Conclusions ................................................................................................................................................48
3.6
3.7
References ..................................................................................................................................................48
4 BIOLOGICAL AND PHYSICO-CHEMICAL FORMATION OF BIRNESSITE DURING
RIPENING OF MANGANESE REMOVAL FILTERS ........................................................... 55
4.1
Abstract.......................................................................................................................................................56
4.2
Introduction ...............................................................................................................................................56
4.3
4.4
4.4.1
Analytical data pilot .........................................................................................................................59
4.4.2
Ripening time of filter media .........................................................................................................60
4.4.3
Raman spectroscopy ........................................................................................................................61
4.4.4
Electron paramagnetic resonance (EPR) .....................................................................................62
4.4.5
Scanning electron microscopy (SEM) ..........................................................................................65
4.5
Conclusions ................................................................................................................................................66
4.6
4.7
References ..................................................................................................................................................67
5 IDENTIFACTION OF THE BACTERIAL POPULATION IN MANGANESE REMOVAL
FILTERS..................................................................................................................................... 73
5.1
Abstract.......................................................................................................................................................74
5.2
Introduction ...............................................................................................................................................74
5.3
5.3.1
Next generation DNA sequencing ................................................................................................76
5.3.2
qPCR ..................................................................................................................................................77
5.3.3
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDITOF MS) ...........................................................................................................................................78
5.3.4
Fermenter growth test with selected bacteria to produce biological MnOx ..........................78
5.4
5.4.1
Next generation DNA sequencing ................................................................................................79
5.4.2
qPCR ..................................................................................................................................................79
5.4.3
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry biotyper
(MALDI-TOF MS)..........................................................................................................................80
5.4.4
Fermenter growth test with selected bacteria for the biological production of MnOx ........81
5.5
Conclusions ................................................................................................................................................82
xx
5.6
5.7
References ..................................................................................................................................................83
6 REDUCTION OF RIPENING TIME OF FULL SCALE MANGANESE REMOVAL
FILTERS WITH MANGANESE OXIDE COATED MEDIA ................................................. 91
6.1
Abstract.......................................................................................................................................................92
6.2
Introduction ...............................................................................................................................................92
6.3
6.3.1
MOCS and MOCA ..........................................................................................................................93
6.3.2
Physical and chemical properties of MOCS and MOCA ..........................................................94
6.3.3
Batch adsorption experiments .......................................................................................................94
6.3.4
Full scale filter runs..........................................................................................................................94
6.4
Results and discussion ..............................................................................................................................97
6.4.1
Ripening of virgin filter media in reference filters ......................................................................97
6.4.2
Effect of water quality parameters ................................................................................................98
6.4.3
MOCS and MOCA characterisation and batch adsorption experiments ...............................99
6.4.4
Ripening of full scale filters with the addition of MOCS and MOCA layers ...................... 101
6.5
Conclusions ............................................................................................................................................. 102
6.6
Acknowledgements ................................................................................................................................ 103
6.7
References ............................................................................................................................................... 103
7 FACTORS CONTROLLING THE RIPENING OF MANGANESE REMOVAL FILTERS
IN COVENTIONAL AERATION-FILTRATION GROUNDWATER TREATMENT...... 107
8
7.1
Abstract.................................................................................................................................................... 108
7.2
Introduction ............................................................................................................................................ 108
7.3
Materials and Methods .......................................................................................................................... 109
7.4
Results and Discussion.......................................................................................................................... 112
7.4.1
Filter ripening with virgin media and the effect of filter backwashing ................................ 112
7.4.2
Filter ripening with addition of a layer of manganese coated media .................................... 116
7.5
Conclusions ............................................................................................................................................. 118
7.6
Acknowledgements ................................................................................................................................ 119
7.7
References ............................................................................................................................................... 119
GENERAL CONLUSIONS ................................................................................................. 123
8.1
Overall conclusions and perspective ................................................................................................... 124
8.2
Assessment of manganese removal from over 100 groundwater treatment plants .................... 125
8.3
Manganese removal from groundwater: Characterization of filter media coating ...................... 126
8.4
. Biological and physico-chemical formation of Birnessite during ripening of manganese removal
filters ......................................................................................................................................................... 126
8.5
Identification of the bacterial population in manganese removal filters ....................................... 128
8.6
Reduction of ripening time of full scale manganese removal filters with manganese oxide coated
media ........................................................................................................................................................ 129
8.7
.Factors controlling the ripening of manganese removal filters in conventional aeration-filtration
groundwater treatment .......................................................................................................................... 130
8.8
General outlook, limitations and recommendations ........................................................................ 130
xxi
List of abbreviations and symbols ......................................................................................................................... 133
List of publications and presentations .................................................................................................................. 135
Curriculum Vitae Jantinus Bruins .......................................................................................................................... 137
xxii
Manganese removal from groundwater: role of biological and physico-chemical autocatalytic processes
1
Figure: Filter media for manganese removal (clockwise from top) ; MOCS, Virgin Sand, Extruded
Activated Carbon, Glass, Anthracite, IOCS, Zeolite and Limestone
(photo by J.H. Bruins, 2010)
0
1
GENERAL INTRODUCTION
“Access to safe drinking water is essential to health, a basic human right and a component of effective
policy for health protection”
(WHO, 2011)
1
1.1 Manganese in groundwater and groundwater treatment in
The Netherlands
1.1.1 Manganese occurrence in groundwater
Groundwater is an important source for drinking water production in The Netherlands (Vewin, 2012), in
Europe (EU, 2008) and world-wide (UNEP, 2008). Just like iron, manganese is a commonly occurring
contaminant present in most groundwater (WHO, 2004).
In nature, manganese occurs as a compound, found in many types of rocks. It is usually found together with
iron and silica, and is the 10th most abundant element in the earth’s crust (ATSDR, 2008; IMnI, 2010). It is
a constituent of more than 30 manganese oxide/hydroxide based minerals, playing an active role in the
environmental geochemistry (Post, 1999). Manganese oxides are ubiquitous in soils and sediments, and
because they are highly chemically reactive and strong scavengers of heavy metals, they exert considerable
influences on the chemical behaviour of sediments, soils and associated aqueous systems. Manganese can
exist in multiple oxidation states. The environmentally and biologically most important minerals are those
containing Mn2+ or Mn4+ (USEPA, 2004). Due to the natural occurrence of manganese in sediments and
soils, manganese is also present in associated aqueous systems with low redox potential and pH, such as
anaerobic groundwater. The occurrence and thermodynamic stability of different manganese species in
natural waters depends on conditions like redox, pH, temperature and oxygen concentration. The most
abundant and stable manganese species in anaerobic groundwater (low pH and redox potential) is Mn2+
(Stumm and Morgan, 1996).
1.1.2 The relevance of manganese in drinking water and guideline
values
Trace levels of manganese are essential for growth and development of humans, animals and plants.
However, for health and aesthetic considerations the amount of manganese in drinking water should be
limited to very low values.
Health concern associated with manganese presence in drinking water is mostly related to neurologic
symptoms (Santamaria and Sulsky, 2010; Rodriguez et al., 2013,). Manganese-induced clinical neurotoxicity
is also associated with a motor dysfunction syndrome commonly referred to as Manganism (a Parkinsonlike disorder). Dosages of 1 to 150 mg/kg of body weight per day (‘short term exposure’) and 1 to 2 mg/kg
body weight per day (‘long term exposure’) (WHO, 2004) of oral or inhalation exposure are associated with
increased manganese levels in tissue. That may lead to the development of these adverse neurological,
reproductive, or respiratory diseases. However, an increasing number of studies report associations between
neurologic symptoms and manganese exposure in infants and children. These findings, in combination with
the questionable scientific background of data used in setting the current manganese WHO-guideline value
for drinking water at 400 µg/L, warrant re-evaluation of the guideline (WHO, 2004; Ljung and
Vather, 2007). Especially the manganese uptake through drinking water consumption by babies could be
relatively high due to their low body weight, relatively high intake and relatively poor manganese excretion
(Brown and Foos, 2009). In addition, the presence of manganese exacerbates the health problems caused
by arsenic (Bunderson et al., 2006, Wright et al., 2006). Thus, the presence of manganese in drinking water
is of particular risk in those areas that also contain elevated levels of arsenic. Therefore the WHO-health
based guideline value for manganese in drinking water is 400 µg/L (WHO, 2011). In the US, the lifetime
health advisory value is 300 µg/L and the EPA’s Secondary Maximum Contaminant Level is 50 µg/L. The
former WHO guideline value of 100 µg/L is still valid in several countries (e.g., Serbia and Jordan) while the
2
guideline value of the EU is 50 µg/L (EU, 1998). This is also the guideline value for manganese according
to the Dutch Revised Drinking Water Decree (VROM, 2011). The internal standard for several Dutch and
Belgian Water Supply Companies varies from 5 to 20 µg/L based on aesthetic considerations and practical
problems in drinking water treatment.
1.1.3 Groundwater treatment in the Netherlands
In The Netherlands, 77 % of the groundwater sources, contain manganese (Mn2+) in concentrations higher
than the EU and Dutch (NL) guideline value of 50 µg/L, whereas approximately 14 % contains more Mn2+
than the health related WHO guideline value of 400 µg/L (Fig. 1.1).
Manganese (µg/L)
400
300
200
100
0
0
20
40
60
80
100
Locations (%)
WHO (400 ug/L): < 86%
EC/DUTCH (50 ug/L): <23%
all locations
Figure 1.1: overview of manganese concentrations in Dutch groundwater (2013), together with the
WHO, EU and Dutch (NL) guideline values for manganese in drinking water (RIVM, 2013)
In The Netherlands, manganese is removed from groundwater by aeration-rapid (sand) filtration (Fig. 1.2).
Such a treatment process is easy to operate, cost effective and sustainable, as does require use of strong
oxidants (e.g., O3, Cl2, ClO2, KMnO4), with the associated risk of disinfection by-product formation and
over or under dosing. Also this process does not require use of special adsorbents
(e.g., manganese green sand), which are applied in other countries (e.g., USA) and which must be regenerated
in time (Knocke et al., 1991). However, the application of aeration-filtration in practice is not completely
problem-free.
3
Figure 1.2 : Typical aeration-filtration groundwater treatment scheme with pre aeration, (down flow)
first rapid filtration step (sand), second aeration and second rapid filtration step (sand) (source: Water
Supply Company Groningen)
1.2 Problem description regarding manganese removal in
practice
In Western European countries such as The Netherlands, Germany and Belgium, problems caused by
manganese are related to aesthetic considerations and practical problems in drinking water treatment, rather
than health related problems. Aesthetic considerations are organoleptic properties like undesirable taste,
staining of plumbing fixtures and laundry (WHO, 2004) and deposition of manganese oxides in distributions
systems, causing black water incidents (Sly et al., 1990). Practical problems in drinking water treatment
associated with manganese presence, express themselves in blocking of filter nozzles and clogging of filters,
valves and piping.
To remove manganese from groundwater, aeration followed by rapid sand filtration, is commonly applied.
However, manganese removal by aeration-filtration is frequently associated with problems, such as:
•
•
•
very long ripening periods (several months to more than one year) are required to achieve an
effective removal with new filter media, leading to associated costs and reduced production capacity
(Cools, 2010; Huysman, 2010; Krull, 2010);
frequent manganese breakthrough of filters after some years of operation, resulting in filter media
replacement. This phenomenon is again associated with long start-up and additional costs for filter
media disposal and replacement (De Ridder, 2008);
inefficient manganese removal from some groundwater types, e.g., with a low pH (< 7) , and a high
concentration of ammonia (NH4 > 2.5 mg/L) (Gouzinis et. al., 1988).
Particularly, the long ripening time of filter media is a major concern and is subject of this study. Due to the
long ripening time, water companies have to waste large volumes of treated water, making this process less
sustainable. In 2013, the Water Company Groningen had to waste more than 100,000 m3 of pre treated
water during start up of one filter with virgin filter media. In addition, costs associated with filter media
ripening (man power, electricity, operational and analysis costs) are high. Furthermore, a filter in the process
of ripening cannot be used for the production of drinking water. Decreasing filter ripening time for
4
manganese removal is a serious concern for water supply companies. Several mechanisms are suggested to
be involved in the startup of manganese removal filters with virgin media, and include the following:
•
•
•
physical and chemical processes, e.g., adsorption, oxidation, (co)precipitation;
microbiological manganese oxidation and removal;
combinations of the aforementioned processes.
Despite the fact that already a lot of research has been done on manganese removal, the controlling
mechanisms, especially of the startup of filter media ripening, are not yet fully understood. Several research
publications suggest that ripening time in manganese removal by conventional aeration-filtration could be
lowered by introducing manganese and/or iron (hydro) oxide coated filter media to conventional rapid sand
filters (Buamah, 2009; Buamah et al., 2008, 2009a; Hu et al. 2004a,b; Islam et al., 2010; Kim et al., 2009;
Olanczuk-Neyman et al., 2000; Sahabi et al., 2009; Stembal et al., 2005; Tekerlekopoulou and Vayenas, 2008;
Tiwari et al., 2007). Results indicated that oxidation of manganese and iron, adsorbed on filter media during
groundwater treatment play an important role in good long-term performance of manganese removal filters
(Buamah et al., 2009). Furthermore, results obtained by Buamah (2009) showed that water quality (e.g., pH
and HCO3-) also plays an important role in manganese removal by influencing the solubility of Mn2+
(Fig. 1.3).
5.0
Mn (II) solubilities (mg/L)
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0
100 200 300 400 500 600 700 800 900 1000
HCO3- (mg/L)
pH 6
pH 6.5
pH 7
pH 7.5
pH 8
Figure 1.3 : solubility of Manganese (Mn 2 + ) as a function of pH and HCO 3 - concentration (adopted
from Buamah, 2009)
It is also known from literature that biology may play an important role with respect to manganese removal.
Several researchers investigated the influence and the capability of different types of bacteria to oxidize
manganese, amongst others Leptothrix sp. (Adams and Ghiorse, 1985; Barger et al., 2009; Boogerd and De
Vrind, 1987; Burger et al., 2008a; Burger et al., 2008b; Corstjens et al., 1997; El Gheriany et al. 2009; Hope
and Bott, 2004; Tebo et al. 2004, 2005); Pseudomonas sp. (Barger et al., 2009; Brouwers et al., 1999; Caspi et al.,
5
1998; DePalma, 1993; Gounot, 1994; Tebo et al., 2004, 2005; Villalobos et al., 2003, 2006) and Bacillus sp.
(Barger et al., 2005, 2009; Brouwers et al., 2000; Mann et al., 1988; Tebo et al., 2004, 2005). However, the
presence of bacteria that are potentially able to oxidize manganese, is not a guarantee of substantial
manganese oxidation and subsequent manganese removal in filters.
Consequently, additional research is still required. In particular, the type of MnOx present in filter media
coatings needs to be identified, how this manganese oxide coating is formed on virgin filter media and the
role of biology in the process. Also the manganese adsorption capacity of manganese coated filter media,
such as manganese oxide coated sand (MOCS) and manganese oxide coated anthracite (MOCA) has to be
investigated, as well as their role in the adsorption and subsequent oxidation of manganese. Furthermore,
knowledge of parameters playing an important role in manganese removal such as water quality, operational
and design parameters is limited, and should be investigated in more detail if the process is to be optimized.
Therefore, additional information is required on:
•
•
•
•
•
•
key factors required for efficient manganese removal in practice, in conventional aeration filtration
ground water treatment plants (GWTPs);
the effect of raw water quality parameters (e.g., pH, Fe2+, NH4+) on manganese removal;
the type of MnOx present in filter media coatings, like MOCS and MOCA;
the mechanisms responsible for a rapid start up of virgin filter media ripening;
the type of MnOx formed initially on virgin filter media;
the contribution of biological processes in the start up of manganese removal.
1.3 Aim and research objectives of this thesis
The main goal of this PhD research is to get a better understanding of the mechanisms involved in the
ripening of virgin filter media used in manganese removal. Furthermore, this research will allow the
development of an innovative manganese removal process that can shorten or completely eliminate ripening
in new filters, as well as prolonging the lifetime of filter media.
In order to achieve the established goals, the following research objectives were defined:
1. Prepare an overview of manganese removal in selected aeration filtration groundwater treatment
plants, including groundwater quality, process conditions applied and operational experiences.
2. Characterise and identify the manganese oxide(s) present in filter media coatings in conventional
aeration-filtration groundwater treatment plants, with efficient manganese removal.
3. Provide a better understanding of the processes involved in ripening virgin filter media by
performing ripening experiments at pilot scale, and propose measures to speed up filter media
ripening and subsequent manganese removal.
4. Investigate the role of biological processes in manganese removal on laboratory, at pilot (batch)
and full scale and propose beneficial conditions for manganese oxidizing bacteria.
5. Investigate the potential of filter media with a high manganese oxide content (e.g., MOCS and
MOCA) to accelerate filter media ripening in conventional aeration filtration filters in practice, and
propose measures resulting in improved manganese removal.
6
6. Improve understanding of process and operational conditions applied in practice (e.g., back
washing, iron load) in conventional aeration-filtration systems and identify key factors to optimize
manganese removal in practice.
1.4 Outline of this thesis
The introduction of this thesis (chapter 1) describes the problem statement as well as general information
about groundwater treatment and the need for manganese removal, based on health, aesthetic and
operational aspects. This chapter ends with the goal of this study, the research objectives and the thesis
outline.
Chapter 2 deals with water quality and operational parameters that affect manganese removal in aeration rapid sand filtration systems. Data from selected full-scale groundwater treatment plants in The Netherlands,
Belgium, Germany, Jordan and Serbia were collected. The focus of the overview was to assess the effect of
groundwater quality, process design and operational parameters on manganese removal efficiency in the
first aeration-filtration stage of treatment plants. In the first filtration stage, manganese is removed together
with iron and ammonia. Single and multiple statistical relations between manganese removal and water
quality data, process design and operational parameters were conducted to define key factors responsible
for manganese removal in practice.
Chapter 3 covers the characterization and identification of naturally formed manganese oxide coatings that
can remove Mn2+ in conventional groundwater treatment plants (GWTPs). Characterization was carried out
via X-ray diffraction (XRD), Scanning Electron Microscopy (SEM) and Raman spectroscopy analyses.
Furthermore based on Electron Paramagnetic Resonance (EPR) analysis, the biological or physico-chemical
origin of the manganese oxide present in the filter media coating was established.
Chapter 4 describes the ripening process of virgin media in a pilot under well defined and controlled
conditions. This media ripening process was followed for more than 600 days. During this study, manganese
oxide on filter sand and in the backwash water was investigated by the characterization and identification
techniques mentioned in chapter 3. The origin of the formed manganese oxide at the start of filter media
ripening and during prolonged filter media ripening was determined. Based on the evolution of manganese
oxide formed during ripening, the mechanisms involved were explained.
Chapter 5 deals with the role of biology in filter ripening and subsequent manganese removal. In this study
advanced molecular techniques, such as qPCR, DNA pyro sequencing and Malditof protein analyses were
employed throughout the pilot research phase mentioned in chapter 4. This study also yields information
on bacteria species present during ripening. Furthermore in this chapter the conditions beneficial for
biological manganese removal are described.
Chapter 6 describes the results of filter ripening in practice and the influence of employing coated filter
media. The filter media ripening with and without coated media were investigated at two full scale
groundwater treatment plants. Different virgin filter media (anthracite and sand) were employed at each
location and Manganese oxide coated sand (MOCS) was added in one location and manganese oxide coated
anthracite (MOCA) at the other location. The aim of this study was to get information about the adsorptive
capacity of the two filter media coatings and their effect on filter media ripening in practice.
Chapter 7 recommends some key factors to speed up filter media ripening in practice. Therefore, the
benefits of using coatings and the differences between the use of MOCS and MOCA in practice (chapter 6)
7
are further investigated in a pilot research (at GWTP Grobbendonk). In this pilot research, the focus was
to investigate the influence of: (1) back washing (or filter loading with particles), (2) the differences between
selected virgin media (anthracite and sand) on filter ripening, (3) differences in the ripening times between
virgin filter media and MOCA and MOCS and (4) the use of dry and fresh coated filter media. Based on the
outcome and results of this study, a strategy to optimize ripening is outlined.
Finally, chapter 8 summarizes the main results, conclusions and limitations of this research. Also
recommendations for practical use and suggestions for follow up research are given.
1.5 References
Adams, L.F., Ghiorse, W.C.1985. Influence of Manganese on Growth of a Sheathless Strain of
Leptothrix discophora. Aplied and environmental Microbiology, 49 (3), 556-562.
ATSDR. 2008. (draft, update 2000) Toxicological profile for manganese, U.S. Department of Health and
Human Services, Agency for Toxic Substances and Disease Registry, Atlanta.
Barger J.R., Tebo, B.M., Bergmann, U., Webb, S.M., Glatzel, P., Chiu, V.Q., Villalobos, M. 2005. Biotic and
abiotic products of Mn(II) oxidation by spores of the marine Bacillus sp. strain SG-1. American
Mineralogist, 90, 143-154.
Barger, J.R., Fuller. C.C., Marcu. M.A., Brearly A., Perez De la Rosa M., Webb S.M., Caldwell W.A. 2009.
Structural characterization of terrestrial microbial Mn oxides from Pinal Ckeek, AZ. Ceochimica et
Cosmochimica Acta 73, 889-910.
Boogerd, F.C., De Vrind J.P.M. 1987. Manganese oxidation by Leptothrix discophora, Journal of bacteriology,
489-494.
Brouwers, G.J., Vrind de J.P.M, Corstjens P.L.A.M., Cornelis P, Baysse C, Vrind de Jong E.W. 1999. CumA,
aGene Encoding a multicopper oxidase, is involved in Mn2+oxidation in Pseudomonas putida GB-1.
Applied and environmental microbiology , 65 (4), 1762-1768.
Brouwers G.J., Vijgenboom E., Corstjens P.L.A.M., de Vrind J.P.M., de Vrind-de Jong E.W. 2000. Bacterial
Mn2+ oxidizing systems and multicopper oxidases: an overview of Mechanisms and Functions,
Geomicrobiology Journal, 17, 1-24.
Brown M.T., Foos B. 2009. Assessing Children’s exposure and risks to drinking water contaminants: A
Manganese case study, Human and Ecological Risk Assessment, 15, 923-947.
Buamah, R., Petrusevski, B., Schippers, J.C. 2008. Adsorptive removal of manganese (II) from the aqueous
phase using iron oxide coated sand. Journal of Water supply: Research and Technology-AQUA 57.1,
1-11.
Buamah, R. 2009. Adsorptive removal of manganese, arsenic and iron from groundwater. PhD-thesis,
UNESCO-IHE Delft / Wageningen University, The Netherlands.
8
Buamah, R., Petrusevski, B., de Ridder, D., van de Watering and Schippers, J.C., 2009a. Manganese removal
in groundwater treatment: practice, problems and probable solutions. Journal of Water Science and
Technology: Water Supply 9.1, 89-98.
Bunderson M., Pereira F., Schneider M.C., Shaw P.K., Coffin J.D., Beall H.D. 2006. Manganese enhances
peroxynitrite and leukotriene E4 formation in bovine aortic endothelial cells exposed to arsenic,
Cardiovasc. Toxicol., 6 (1), 15-23.
Burger, M.S., Krentz, C.A., Mercer, S.S., Gagnon, G.A. 2008a. Manganese removal and occurrence of
manganese oxidizing bacteria in full-scale biofilters. Journal of Water Supply: Research and
technology-AQUA 57.5, 351-357.
Burger, M.S., Mercer, S.S., Shupe, G.D., Gagnon, G.A. 2008b. Manganese removal during bench-scale
biofiltration, Water Research 42, 4733-4742.
Caspi, R., Tebo, B.M., Haygood, M.G. (1998). C-Type Cytochromes and Manganese Oxidation in
Pseudomonas putida MnB1. Applied and environmental Microbiology, 64.10, 3549-3555.
Cools B. 2010. De Watergroep, personal communication, Belgium.
Corstjens, P.L.A.M., de Vrind, J.P.M., Goosen, T., de Vrind-de Jong, E.W. 1997. Identification
and molecular analysis of the Leptothrix discophora SS-1 mofA gene, a gene putatively encoding a
manganese-oxidizing protein with copper domains. Geomicrobiology Journal, 14 (2), 91-108.
DePalma, S.R. 1993. Manganese oxidation by Pseudomonas putida, PhD-thesis, Harvard University,
Cambridge, Massachusetts, USA.
El Gheriany I.A., Bocioaga B., Anthony Hay A.D., Ghiorse W.C., Shuler M.L., Lion L.W. 2009. Iron
Requirement for Mn(II) Oxidation by Leptothrix discophora SS-1, Applied and Environmental
Microbiology, 75 (5),1229-1235.
EU. 1998. Drinking Water Directive, Council directive 98/83/EC.
EU. 2008. Groundwater protection in Europe - The new groundwater directive, European commission.
Gounot A-M. 1994. Microbial oxidation and reduction of manganese: Consequences in groundwater and
applications, FEMS Microbiology reviews 14, 339-350.
Gouzinis, A., Kosmidis N., Vayenas D.V., Lyberatos G. 1988. Removal of Mn and simultaneous removal
of NH3, Fe and Mn from potable water using a trickling filter, Wat. Res. 32 (8), 2442-2450.
Hope C.K., Bott T.R. 2004. Laboratory modelling of manganese biofiltration using biofilms of Leptothrix
doscophora, Water Research 38, 1853-1861.
Hu, P-Y., Hsieh, Y-H., Chen, J-C., Chang, C-Y. 2004a. Adsorption of divalent manganese ion on
manganese-coated sand. Journal of Waters Supply: Research and Technology-AQUA 53.3, 151-158.
Hu, P-Y., Hsieh, Y-H., Chen, J-C., Chang, C-Y. 2004b. Characteristics of manganese-coated sand using
SEM and EDAX analysis. Journal of Colloid and interface science, 272, 308-313.
9
Huysman K. 2010. fPidpa Department of Process Technology and Water Quality, personal communication,
Belgium.
IMnI. 2010. The international Manganese Institute, Paris, France, http://www.manganese.org.
Islam A.A., Goodwill, J.E., Bouchard, R., Tobiasen, J.E., Knocke, W.R. 2010. Characterization of filter
media MnO2(s) surfaces and Mn removal capability. Journal AWWA, 102 (9), 71-83.
Kim, W.G., Kim, S.J., Lee, S.M., Tiwari, D. 2009. Removal characteristics of manganese-coated solid
samples for Mn(II). Desalination and water treatment 4, 218-223.
Knocke, W.R., Van Benschoten, J.E., Kearny, M.J., Soborski, A.W., Reckhow, D.A. 1991.
Kinetics of Manganese and Iron oxidation by Potassium Permanganate and Chlorine dioxide, Journal of
AWWA, June 1991, 80-87.
Krull J., Stadtwerke Emden (SWE). 2010 Personal communication, Germany.
Ljung K., Vahter M. 2007. Time to Re-evaluate the guideline value for manganese in Drinking water ?,
Environmental Health Perspectives, 115 (11), 1533-1538.
Mann, S., Sparks. N.H.C., Scoot G.H.E., Vrind- De Jong E.W. 1988. Oxidation of Manganese and
formation of Mn3O4 (Hausmannite) by spore Coats of a marine Bacillus sp. Applied and
environmental microbiology, 54 (8), 2140-2143.
Olanczuk-Neyman, K., Bray, R. 2000. The role of Physico-Chemical and Biological Processes in Manganese
and Ammonia Nitrogen Removal from Groundwater. Polish Journal of Environmental studies, 9 (2),
91-96.
Post, J. E. 1999. Manganese oxide minerals: Crystal structures and economic and environmental significance,
Proceedings of the National Academy of Sciences USA, 96, 447-3454.
Ridder de D. 2008. Opstart mangaanverwijdering in snelfilters, BTO 2008.014 (in Dutch).
RIVM - National Institute for Public Health and the Environment. 2013. Information groundwater
composition in The Netherlands 2012, provided by email, Mr. Dik.
Rodríguez-Barranco M., Lacasaña M., Aguilar-Garduño C., Alguacil J., Gil F., González-Alzaga B., RojasGarcía A. 2013. Association of arsenic, cadmium and manganese exposure with neurodevelopment
and behavioral disorders in children: A systematic review and meta-analysis. Science of the Total
Environment 454-455, 562-577.
Sahabi, D.M., Takeda M., Suzuki I., Koizumi J-I. 2009. Removal of Mn2+ from water by “aged” biofilter
media: The role of catalytic oxides layers. Journal of Bioscience and Bioengineering, 107 (2), 151-157.
Santamaria, A. B., Sulsky S.I. 2010. Risk assessment of an essential element: Manganese, Journal of
Toxicology and Environmental Health, part A, 73, 128-155.
10
Sly L.I., Hodgkinson M.C., Vulapa Arunpairojana. 1990. Deposition of manganese in a drinking water
distribution system. Applied and environmental biotechnology, 56 (3), 628-639.
Stembal, T., Markic, M., Ribicic, N., Briski, F., Sipos, L. 2005. Removal of ammonia, iron and manganese
from ground waters of Northern Croatia – pilot plant studies, Process Biochemistry 40, 327-335.
Stumm, W. and Morgan, J.J. 1996. Aquatic Chemistry, chemical equilibria and rates, 3rd ed. Wiley, New
York.
Tebo, B.M., Marger J.R., Clement B.G., Dick G.J., Murray K.J., Parker. D., Verity R., Webb S.M. 2004.
Biogenic Manganese oxides: Properties and mechanisms of formation. Annu. Rev. Earth Planet Sci.,
32, 287-328.
Tebo, B.M., Johnson H.A., McCarthy J.K., Templeton A.S. 2005. Geomicrobiology of manganese (II)
Oxidation. TRENDS in Microbiology, 13 (9), 421-428.
Tekerlekopoulou, A.G. & Vayenas D.V. 2008. Simultaneous biological removal of ammonia, Iron and
manganese from potable water using a trickling filter. Biochemical Engineering Journal 39, 215-220.
Tiwari, D., Yu, M.R., Kim, M.N., Lee, S.M., Kwon, O.H., Choi, K.M., Lim, G.J., Yang, J.K. 2007. Potential
application of manganese coated sand in the removal of Mn (II) from aqueous solutions. Water
Science & Technology, 56 (7), 153-160.
UNEP. 2008. Vital water graphics, http://www.unep.org/dewa/vitalwater/index.html.
USEPA. 2004. Drinking water Health Advisory for Manganese.
VEWIN. 2012 Drinkwaterstatistiek 2010, de watercyclus van bron tot tap (in Dutch).
Villalobos, M., Toner, B., Bargar, J., Sposito, G. 2003. Characterization of the manganese oxide produced
by Pseudomonas putida strain MnB1. Geochimica et Cosmochimica Acta, 67 (14), 2649-2662.
Villalobos, M., Lanson, B., Manceau, A., Toner, B., Sposito, G. 2006. Structural model for the biogenic Mn
oxide produced by Pseudomonas putida. American Mineralogist, 91, 489-502.
VROM. 2011. Ministry of Housing, Spatial Planning and the Environment, Waterleidingbesluit 2011 (in
Dutch).
WHO. 2004. Manganese in Drinking water – Background document for development of WHO Guidelines
for drinking water quality, Geneva, WHO/SDE/WSH/03.04/104.
WHO. 2011. Guidelines for drinking water quality – 4th edition. WHO Press, World Health Organization,
20 Avenue Appia, 1211 Geneva 27, Switzerland.
Wright R.O., Amarasiriwardena C., Alan D. Woolf A.D., Jim R., Bellinger D.C. 2006. Neuropsychological
correlates of hair arsenic, manganese, and cadmium levels in school-age children residing near a
hazardous waste site, NeuroToxicology, 27 (2), 210-216.
11
Figure: Typical aeration-filtration groundwater treatment scheme (GWTP De Punt) with pre
aeration, (down flow) single rapid filtration step (sand) and post aeration (source: Water Supply
Company Groningen)
12
2
ASSESSMENT OF MANGANESE REMOVAL FROM
OVER 100 GROUNDWATER TREATMENT PLANTS
Main part of this chapter was published as:
Jantinus H. Bruins, Dirk Vries, Branislav Petrusevski, Yness M. Slokar, Maria D. Kennedy (2014).
Assessment of manganese removal from over 100 groundwater treatment plants. Journal of Water
Supply: Research and Technology – AQUA 63.4:268 - 280
13
2.1 Abstract
The aim of this study was to make an inventory of water quality and operational parameters which could
affect manganese removal through aeration and rapid sand filtration and to establish correlations between
these parameters and manganese removal efficiency. The focus of the overview was on manganese removal
efficiency in the first aeration-filtration stage of conventional groundwater treatment plants. Data from
selected full-scale groundwater treatment plants have been collected, and univariate and multivariate
statistical analysis were conducted. Multivariate statistics indicated that multiple parameters including NH4+
removal efficiency, iron loading per filter run and pH of filtrate play a significant role in manganese removal,
while other parameters (oxygen concentration in filtrate, filtration rate and empty bed contact time (EBCT))
were found to be of the secondary importance. Univariate statistical assessment of the data suggests that
very effective manganese removal can be achieved when all of the following conditions are met: NH4+
removal efficiency > 85%, pH of filtrate > 7.1, iron loading per filter run < 2.7 kg Fe/m2, oxygen
concentration in filtrate > 1 mg O2/L, filtration rate < 10.5 m/h and EBCT > 11.5 min.
Keywords: Aeration- filtration, conventional groundwater treatment plants, groundwater quality,
inventory, manganese removal efficiency, operational parameters
2.2 Introduction
Groundwater is the predominant global source of water for drinking water production (UNEP 2000). In
addition to impurities such as iron, ammonia and methane, groundwater frequently contains elevated levels
of dissolved manganese, which need to be lowered for both health and aesthetic reasons. Manganese is a
naturally occurring metal that is a constituent of more than 30 manganese oxide/hydroxide minerals, which
occur in a wide variety of geological settings (Post, 1999). The occurrence and thermodynamic stability of
different manganese species in groundwater’s is dependent on different conditions like reduction-oxidation
(redox) potential (Eh) or electron activity (pε) and pH (Fig. 2.1).
14
Figure 2.1: Electron Activity (pε) and redox potential (E h ) – pH diagram for aqueous manganese
(Stumm and Morgan, 1996)
Due to the natural occurrence of manganese in sediments and soils, dissolved manganese (Mn2+) can also
be present in associated aqueous systems like groundwater, because of the low redox potential of anaerobic
groundwater and the relative low pH. From Fig. 2.1 it can be seen that in a reducing environment and
relatively low pH (conditions found in anaerobic groundwater), the most abundant and stable manganese
species is Mn2+.
The removal of manganese from groundwater is commonly achieved through aeration-rapid sand filtration.
A typical example of a groundwater treatment plant (single filtration step with pre- and post-aeration) is
shown in Fig. 2.2. An aeration followed by a single filtration step is often applied, but in some cases, e.g.,
groundwater with elevated levels of ammonia, iron and/or manganese, two filtration steps are required. In
those cases, a second aeration step is frequently included to correct the pH and increase the oxygen
concentration in the (treated) water. Physical-chemical and/or biological mechanisms were reported to be
responsible for manganese removal (Vandenabeele et al., 1995; Graveland & Heertjes, 1975). Despite of
extensive research carried out on manganese removal through aeration-rapid sand filtration, the mechanisms
controlling the process are still not fully understood.
15
Figure 2.2: Typical conventional groundwater treatment scheme with pre aeration, (down flow) single
rapid filtration step and post aeration (source: Water Supply Company Groningen)
Conventional groundwater treatment is simple and very cost effective, but, performance of an individual
plant depends amongst other factors on individual approach to various technological issues related to a
given groundwater. Advantages of manganese removal on conventional aeration-rapid sand filtration could
be seriously hampered by:
(a) long ripening periods taking several weeks to more than a year to achieve effective manganese removal
with virgin filter media (Huysman, 2010;).
(b) manganese breakthrough after some years of operation, introducing the need for filter media
replacement, associated with a long start-up period and additional costs for filter media disposal and
replacement (Buamah et al., 2008). As an example in Fig. 2.3 the manganese breakthrough is expressed after
more than 12 years of proper operating and treating the same raw water quality.
16
c(Mn) in filtrate (mg/L)
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
15-Jan
11-Oct
8-Jul
3-Apr
Old filter media
Figure 2.3: An example of manganese breakthrough in a Dutch groundwater treatment plant based
on aeration-rapid sand filtration (source: Water Supply Company Groningen).
(c) insufficient manganese removal for some types of groundwater (e.g., groundwater containing high
concentration of Natural Organic Matter (NOM), that could cause complexation of Mn2+).
Several researchers reported insufficient manganese removal due to water quality matrix and inappropriate
process design and operational parameters. Amongst others, high concentrations of NH4+ (> 2 mg/L) can
have a negative influence on manganese removal (Gouzinis et al., 1998). In addition a low pH also negatively
affect manganese adsorption (Buamah, 2009) and manganese oxidation (Stumm & Morgan, 1996). Without
a catalyst, homogeneous manganese oxidation and subsequently removal by filtration is only achieved when
pH is at least 8.6 (Graveland, 1971). Finally, dissolved iron present in groundwater can affect manganese
removal, because it can compete with Mn2+ for adsorption sites (Hu et al., 2004a, b). Therefore the main
focus of this study was to improve the understanding of manganese removal in conventional groundwater
treatment plants based on aeration and (down flow) rapid sand filtration. Specifically effect of groundwater
quality, process design and operational parameters on manganese removal was assessed.
2.3 Materials and Methods
Data presented in this paper are based on information gathered from over 100 groundwater treatment plants
(GWTPs) from The Netherlands (65), Belgium (34), Germany (6), Jordan (1) and Serbia (1). The data was
mainly collected through questionnaire, or by visiting selected plants and interviewing key employees that
have extensive operational experience. The focus of the inventory was on the plants that have a (down flow)
single aeration- filtration treatment (with simultaneously removal of manganese, iron and ammonia) or on
the 1st filtration step for plants that have multiple aeration-filtration steps. For GWTPs without a single
aeration-filtration but with multiple aeration-filtration steps, questions were aimed at the first filtration step.
Collected (water quality and process design) parameters where based on their importance for manganese
removal, as suggested in literature (Buamah, 2009; Graveland, 1971; Gouzinis, 1998) but also depending on
data availability (e.g., redox potential is known to be of importance, but is seldom measured on regular base).
Most of these plants included in the inventory have silica as filter media, in some cases in combination with
17
anthracite as double media filter. Plants run for at least one year, so containing “aged” (bio)layers (e.g., Fe
and Mn oxides). The following data was collected or calculated.
Parameters with respect to the quality of anaerobic groundwater:
- concentration of ammonia, iron, manganese, hydrogen carbonate, calcium, phosphate, silica,
methane, pH, dissolved or total organic carbon (DOC/TOC).
Parameters with respect to the water quality after the 1st filtration step:
- concentration of ammonia, iron, manganese, hydrogen carbonate, calcium, oxygen, pH;
- removal efficiency of manganese, ammonia and iron.
Parameters with respect to applied process design:
- iron and manganese loading (kg Fe and Mn/m2/filter run);
- backwash (BW) criteria (head loss, time, volume);
- volume of filtrate produced between two consecutive backwashing cycles (m3);
- filtration rate (m3/m2/h);
- empty bed contact time EBCT (minutes) (tEB in PCA analysis, Fig. 2.4);
- filter area (m2);
- filter bed depth (m) (hFB in PCA analysis, Fig. 2.4);
- flow (m3/h);
- filter configuration (gravity/pressure), and type of filtration (‘dry’ or ‘wet’, with water level below
or above filter media, respectively)
Selected raw and treated water quality and plant design parameters were correlated to the performance of
full-scale plants in terms of Mn2+ removal. The focus was on the first filtration stage, in which Mn2+ removal
was often combined with the removal of ammonia, iron and sometimes methane. Principal component
analysis (PCA) and single correlation analysis of Mn2+ removal efficiency as a function of selected water
quality parameters and process design parameters were conducted, to determine the parameters of
importance. For PCA, data were standardized by subtracting the data values by the process variable means,
and dividing the resulting data values by the standard deviation of the corresponding variable. To check
whether the data set is useful for PCA, Kaiser-Meyer-Olkin (KMO) (Cerny et al., 1977) and the Bartlett’s
tests (Bartlett, 1937) were performed. The KMO statistic assesses whether there is an underlying (latent)
structure in the data. Low values of the KMO statistic, i.e., smaller than an index of 0.6, indicate that the
correlations between pairs of variables cannot be explained by other variables, and that PCA or factor
analysis may not be appropriate. The Bartlett’s test (Bartlett, 1937) is used to evaluate the homoscedasticity
(equal variances across samples) of the data. A limit of 0.30 for the absolute value of the coefficient loading
of a principal component is chosen to distinguish between moderate to high and low correlation.
2.4 Results and Discussion
2.4.1 Multivariate statistics and univariate correlations
Table 2.1 gives an overview of selected design and water quality parameters and correlation coefficients
against manganese removal efficiency by a univariate and multivariate statistical method (PCA) .The KMO
measure of sampling adequacy shows that the data set yields a degree of common variance of 0.65, which
can be considered as sufficient for PCA analysis. Furthermore, Bartlett’s test for dimensionality reveals that
all the variables are necessary to explain non-random variations, in other words, they are correlated.
18
The first two components of the PCA are shown on the left of Fig. 2.4, while the number of components
that account for the explained variance is shown on the right. In general, vectors pointing in the same
direction mean that the associated factors are (positively) correlated, while if pointing away from each other,
an inversely proportional relation exists. Additionally, loading vectors at right angles to each other indicate
either a negligible or no interdependence. The variance of the variables in the data set is by definition of
PCA covered by a summation of the principal components together with their loading coefficients. The
part of variance that is covered by principal components, is called the explained variance. The first
component explains the largest part of the variance in the data.
Table 2.1: Linear correlation coefficients and PCA loading coefficients between manganese removal
efficiency and selected water quality and process parameters (PCA parameters with a value above ±0.30
are shown in bold face). PCA has been performed on standardized data.
Parameter /concentration
Parameter
Manganese
removal
Process
conditions
Raw water
Filtrate
Correlation
with Mn
removal
1st component
Unit
0.337
%
Filtration rate
Bed height
EBCT
Fe loading
Mn2+
Fe2+
NH4+
HCO3Ca2+
PO43SiO2
DOC/TOC
NH4+
NH4+ removal
pH
O2
m/h
m
min
kg/m2
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
mg/L
%
mg/L
PCA
(N=34, rank of data set =17)
-0.5046
-0.2859
0.3349
-0.6861
-0.0362
-0.5527
0.0354
0.3167
0.3159
-0.2469
-0.3732
0.0483
-0.2018
0.7618
0.6164
0.0120
-0.218
-0.210
0.167
-0.343
-0.093
-0.316
0.157
0.270
0.262
-0.204
0.112
0.154
-0.268
0.331
0.333
0.063
2nd
component
-0.042
-0.248
-0.265
0.130
-0.014
-0.497
0.032
0.315
0.144
0.026
0.364
0.013
-0.126
0.312
-0.090
-0.045
-0.474
19
Figure 2.4: PCA loadings shown as vectors in the principal component space (left) and number of
components that account for the explained variance of the total data set (right).
Based on both PCA and single correlation analysis (Table 2.1 and Fig. 2.4), NH4+ removal efficiency, iron
loading per filter run, and pH strongly correlate with manganese removal efficiency. In contrast to NH4+
removal efficiency and pH, the iron loading (and often also the iron concentration in raw water) was found
to be inversely proportional to manganese and NH4+ removal. Concentrations of dissolved calcium and
hydrogen carbonate in raw water appear to have low correlation with manganese removal in single
correlation studies, whereas PCA shows that these parameters could have some influence, as shown in the
loadings of the first component. Since the solubility of compounds is directly dependent on pH, it was
expected that calcium and hydrogen carbonate will be (cor)related with pH, thus influencing the loadings
of the first component. Furthermore, the PCA analysis showed certain degree of correlation between all
water quality and process parameters and manganese removal included in the inventory. This confirmed
that Manganese removal in aeration-filtration treatment process is very complex and influenced by a large
array of parameters. From the right hand side of Fig. 2.4 it follows that the second principal component
explains another 15% of the variability in the data, while the third component covers an additional part of
the variance, resulting in approximately 70% coverage of data variance when three components are used.
Hence, the highest PCA loading coefficients of the second component (Table 2.1) indicate that, the influent
manganese concentration, oxygen and phosphate might also play a role in manganese removal. A possible
explanation for the negative (cor)relation between phosphate and manganese removal might be similar to
that for the effect of phosphate on iron removal; heterogeneous Fe2+ oxidation (rate) decreases with
increasing phosphate concentration. The controlling mechanisms are not fully understood yet, but the
oxidation rate may be influenced by colloid formation and their mobility, the charge of filter media, and pH
of (raw) water (O’Melia and Crapps, 1964; Wolthoorn et al., 2004). An additional explanation could be that
phosphate competes with Mn2+ for the same adsorption sites as observed for Fe3+ adsorption on iron hydrooxide surfaces. The PCA further reveals that the variability in data cannot easily be explained by only 2
factors, i.e. at least 4 components are needed to account for approximately 75% of the explained variance
(Fig. 2.4). Although the single correlation analysis does not confirm the effect of oxygen, organic matter
TOC/DOC) and phosphate, it is known from practice that these parameters may also have effect on
manganese removal efficiency. PCA shows that there is indeed an effect, although only significant in the
second principal component for dissolved oxygen (Table 2.1), or even in the third for organic matter. It is
20
important to keep in mind that PCA informs how many parameters are necessary to explain the data and
which variables are most importantly related (to manganese removal), while the following assessment is
based primarily on evaluating the correlation between two variables and analysing the ‘outliers’ by inspection
of the other variables.
In the following sections, the parameters with a predominant effect on manganese removal efficiency are
discussed in more detail.
2.4.2 NH4+ removal efficiency
Fig. 2.5 shows the correlation between manganese and NH4+ removal efficiencies for the treatment plant
included in the inventory. Analysis of the data set (Fig. 2.5) suggests that for a majority of GWTPs, complete
manganese removal is not achieved when NH4+ removal efficiency is less than 85%. A few GWTPs still did
not achieve complete manganese removal even when NH4+ removal was very effective. Closer inspection
reveals that each of the outliers had at least one parameter with values that were unfavourable for effective
manganese removal, i.e., pH in filtrate <7.1 or the filtration rate >10.5 m/h and/or EBCT <11.5 min (both
parameters discussed further in this study).
Mn removal (%)
100
80
60
pH < 7.1
Vf > 10.5 m/h and/or EBCT < 11.5 min
40
20
0
0
10
20
30 40 50 60 70
Removal of NH4 (%)
80
90
100
Figure 2.5: Correlation between manganese and NH 4 + removal efficiencies for full-scale groundwater
treatment plants based on aeration filtration.
The strong correlation between NH4+ and manganese removal, is also observed in the practice. Very similar
trends for NH4+ and manganese removal are specifically observed during the ripening period of new filter
media (Fig. 2.6A). In addition, results from the full scale rapid sand filter suggest that complete manganese
removal occurs only at a certain depth of a filter, when NH4+ removal is almost complete or NH4+ is not
present anymore (Fig.2.6B).
21
Mn and NH4 removal (%)
100
Filter depth (m)
80
60
40
20
0
0
A
5
10
15
20 25 30 35 40 45
Ripening time (days)
ammonium
manganese
50
0.2
0.0
-0.2
-0.4
-0.6
-0.8
-1.0
-1.2
-1.4
-1.6
-1.8
-2.0
-2.2
FILTER
0
55
B
10
20
30
40 50 60 70
Removal (%)
Mn
NH4+
80
90
100
Figure 2.6: Removal efficiency of NH 4 + and manganese during the ripening period of new filter
media of a full-scale groundwater treatment plant (water supply company Groningen, The Netherlands).
A: Manganese and NH 4 + removal efficiency in a single aeration-rapid sand filtration as a function of
the ripening period. B: removal efficiency of manganese and NH 4 + over the depth of a ripened filter.
Several papers confirm that NH4+ removal and manganese removal are related. Scherer and Wichmann
(2000) and Flemming et al. (2004), suggested that a negative effect of NH4+ on manganese removal is a
consequence of poor Mn2+ oxidation due to the low redox potential of the groundwater in the presence of
NH4+. As a result, high concentrations of NH4+ (>2 mg/L) in (raw) groundwater hinder efficient and
complete manganese removal in conventional one-stage aeration-filtration systems Gouzinis et al. (1998).
The observed correlation between NH4+ and manganese removal efficiencies may be also related to
biological processes, i.e., the oxidation of NH4+ by nitrifying bacteria. Several mechanisms may contribute:
-
-
oxidation of Mn2+ by nitrifying bacteria (e.g., Nitrosomonas sp.), although from an energetic point of
view this is unlikely (Vandenabeele et al., 1995);
Adsorption of positively charged dissolved Mn2+ on negatively charged organic compounds, e.g.,
Extracellular Polymeric Substances (EPS) excreted by nitrifying bacteria (Vandenabeele et al., 1995);
Local reduction of the zero point of charge (pHPZC) or Zeta potential of the filter media to more
negative, due to a pH decrease on micro scale, induced by the nitrification reaction; more negatively
charged filter media surface will considerably increase the attraction of positively charged Mn2+.
Increase of the redox potential due to the nitrification reactions resulting in enhanced Mn2+
oxidation.
However additional research is required to further clarify mechanisms responsible for the effect of NH4+
(removal efficiency) on manganese removal.
2.4.3 Effect of iron loading
Iron loading, defined as the amount of iron removed in a filter during one filter run, appears to negatively
affect manganese removal (Table 2.1 and Fig. 2.7).
22
100
80
Mn removal (%)
Mn removal (%)
100
60
40
20
80
60
40
20
0
0
0
2
4
6
8
Fe load
10
12
14
16
18
0
(kg/m2)
5
10
15
20
25
30
35
Fe in raw water (mg/L)
pH < 7.1
NH4+ removal < 85%
Figure 2.7: Manganese removal efficiency as a function of the iron loading per filter run (kg
Fe/m 2 ) and the iron concentration (mg/L) in raw groundwater
From Fig. 2.7, it seems that complete manganese removal cannot be achieved with an iron loading above
approximately 2.7 kg Fe/m2 per filter run. Although iron oxides can adsorb Mn2+, and can consequently
also act as a catalyst for manganese removal (Buamah, 2009), iron (hydr)oxide can also cover active sites on
filter media available for Mn2+ adsorption (Buamah et al., 2009). In addition Fe2+ ions can compete with
Mn2+ ions for adsorption sites (Hu et al., 2004a,b). As a consequence, better manganese removal is typically
achieved when either iron is removed prior to manganese removal, or when iron loading per filter run is
lower. The latter can be achieved by reducing the amount of water filtered between two backwash cycles by
e.g., reducing the filter run length. Data from the conducted inventory also suggest that complete manganese
removal in a single aeration-filtration treatment can be achieved also for groundwater with very high iron
content (up to approximately 15 mg/L) when iron loading per filter run is limited to < 2.7 kg/m2 (Fig. 2.7).
Iron loading can also be influenced by the prevailing iron removal mechanism. The achieved iron loading is
typically much lower when iron is removed through oxidation-floc formation in comparison to adsorptive
and/or biological iron removal. Iron removal through oxidation and floc formation is characterised by
accumulation of iron (hydr)oxide flocs in the upper part of the filter bed or in the anthracite part of a dual
media filters, and associated faster head loss increase, and the need for frequent backwashing, specifically
when fine sand size fraction is used. Consequently, iron removal predominantly through floc formation
results in relatively lower iron loading and has potentially a larger adsorption/oxidation area for Mn2+
adsorption. At the same time more frequent backwashing cycles could negatively influence development of
catalytic layer on the filter media.
A closer look at the inventory data revealed that each of the outliers that showed poor manganese removal
at low iron loadings had at least one other parameter that are critical for manganese removal including:
-
pH < 7.1 (discussed under “effect of (filtrate) pH”) or
NH4+ removal efficiency < 85%, or
filtration rate > 10.5 m/h and/or EBCT < 11.5 min (discussed under “effect of filtration rate” and
“effect of contact time and filter bed depth”).
23
2.4.4 Effect of (filtrate) pH
Both PCA and single correlation analysis of the inventory data show that pH has large effect on manganese
removal efficiency (Fig. 2.8).
Mn removal (%)
100
80
60
NH4+ removal < 85%
Fe load > 2.7 kg/m2.FR
40
20
0
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0
pH
Figure 2.8: Manganese removal efficiency as a function of filtrate pH.
The data presented in Fig. 2.8 show that 7.1 is the limiting pH value, below which complete manganese
removal was not achieved on any of the plants included in the inventory, with the exception of one specific
groundwater treatment plant where complete manganese removal is achieved at a somewhat lower filtrate
pH of 6.8. Closer data inspection of this specific plant revealed that the aerated water had a pH > 7.1, and
that pH was strongly reduced in the filter bed, probably due to the oxidation of Fe2+ (15 mg/L), which
resulted in the release of protons. NH4+ which can also lower pH, is hardly present in raw water (0.1 – 0.25
mg/L). It is very likely that the pH in the upper part of filter bed was around 7.1, which was sufficient for
effective manganese removal, although in this zone iron might compete with manganese for adsorption
sites. More research is, however, needed to explain efficient manganese removal archived at pH <7.1 in this
plant. Also in literature complete (biological) manganese removal at low pH of 6.5 is reported (Burger et al.,
2008). However, it should be mentioned that raw water quality used in the study of Burger et al. (2008)
differs significantly from raw groundwater quality of plants included in this study (extremely low alkalinity,
absence of iron and ammonia, high redox potential). Data shown in Fig. 2.8 strongly indicate that increasing
pH, in general, improves manganese removal. Higher pH will strongly enhance adsorption of Mn2+
(Buamah, 2009), and will support heterogeneous and autocatalytic manganese adsorption and oxidation. At
low pH values, it is known from practice and reported in the literature that physicochemical manganese
removal seems to be impossible in traditional aeration-filtration treatment plants. Ramstedt et al. (2002)
showed that γMnOOH (manganite) starts to dissolve significantly at pH below 6, causing leaching of Mn2+
from the adsorbent into the filtrate. Hastings and Emerson (1986) and Klewicki and Morgan (1999)
described these dissolution processes as disproportion reactions (Eqs. 2.1, 2.2).
Mn3O4
2MnOOH
+
+
2H+
2H+
↔
↔
2MnOOH
MnO2
+
+
Mn2+
Mn2+
+
2H2O
(2.1)
(2.2)
From Fig. 2.8 it is also clear that manganese removal is very effective at pH values between 7.1 and 8.0.
Graveland (1971) and Graveland and Heertjes (1975) reported that there was no substantial oxidation of
Mn2+ with oxygen in water (homogeneous oxidation) at pH < 8.6 when no catalyst is present. Hence, it is
reasonable to expect that the mechanism responsible for manganese removal in treatment plants included
24
in this study is (i) either biological or (ii) a heterogeneous and autocatalytic manganese adsorption and
oxidation that takes place on the surface (e.g., MnOx) of the filter media (Katsoyiannis and Zouboulis, 2004).
Fig. 2.8 depicts that manganese removal was less effective on a number of plants included in the inventory
where the filtrate pH was higher than 7.1. Close inspection of the data for these plants revealed that at least
one of the following parameters limited manganese removal efficiency:
-
NH4+ removal efficiency < 85% or
iron loading per filter run > 2.7 kg Fe / m2 or
filtration rate > 10.5 m/h and/or EBCT < 11.5 min (discussed under “effect of filtration rate” and
“effect of contact time and filter bed depth”).
2.4.5 Other parameters
According to the single correlation analysis (Table 2.1), the rest of the variables included in this study did
not show a significant correlation with the efficiency of manganese removal. However, PCA demonstrates
that some parameters (e.g., filtration rate, empty bed contact time and to a lesser extent oxygen and
phosphate) may influence manganese removal (Table 2.1 and Fig. 2.4). When plotting a number of
parameters such as filtration rate, empty bed contact time (EBCT), and oxygen concentration, against
manganese removal efficiency, some trends may indeed be deduced. The effect of these three parameters is
discussed in the following paragraphs.
2.4.6 Effect of filtration rate (m/h)
In Fig. 2.9, the manganese removal efficiency obtained on the plants included in the study is plotted as a
function of applied filtration rate.
Mn removal (%)
100
80
60
NH4+ removal < 85%
pH < 7.1
EBCT < 11.5 min
40
20
0
0
2
4
6
8
10
12
14
16
18
20
Filtration rate (m/h)
Figure 2.9: Manganese removal efficiency as a function of the filtration rate (m/h).
Based on the results from the inventory, complete manganese removal was achieved only in treatment plants
that operate at filtration rates of up to 10.5 m/h. Ineffective manganese removal was, however, observed at
several treatment plants that operate at filtration rates significantly lower than < 10.5 m/h. These outliers
are most likely due to:
-
poor NH4+ removal efficiency (< 85%) or
high iron loading (> 2.7 kg Fe / m2/filter run) or
low (filtrate) pH (< 7.1) or
25
-
short EBCT (< 11.5 min) (discussed further under “effect of contact time and filter bed depth”).
Lower filtration rates provide longer contact time in the filter bed, and consequently more time for
manganese adsorption and oxidation.
2.4.7 Effect of contact time and filter bed depth
Fig. 2.10 shows manganese removal efficiency as a function of the empty bed contact time (EBCT) for
groundwater treatment plants included in the study. It can be seen that only plants that apply an EBCT of
11.5 minutes or more achieved complete manganese removal.
Mn removal (%)
100
80
60
NH4+ removal < 85%
40
20
pH < 7.1
0
0
5
10 15 20 25 30 35 40 45 50 55 60
Empty bed contact time (min)
Figure 2.10: Manganese removal efficiency as a function of the empty bed contact time (EBCT) in
minutes.
Results from the inventory also show that an EBCT longer than 11.5 minutes does not guarantee complete
manganese removal. Detailed analysis of plants that apply an EBCT > 11.5 minutes and have ineffective
manganese removal showed that again one or more earlier identified water quality parameters and/or
process conditions were probably responsible for poor manganese removal.
A parameter directly related to EBCT is bed height of the filter media. No significant statistical correlation
between bed height and manganese removal efficiency was found (Table 2.1).
Assuming that a groundwater treatment plant is operating at a maximum acceptable filtration rate of 10.5
m/h, a filter bed of approximately 2.0 m is required to provide a minimal required EBCT of 11.5 minutes
that will allow an effective manganese removal. Selected data from the inventory, however, demonstrate
that a relatively shallow filter bed height of 0.9 m might be sufficient to obtain complete Mn2+ removal, if
the other parameters of relevance for manganese removal are optimal (e.g., high pH, low iron loading, low
filtration rate, long EBCT).
2.4.8 Effect of oxygen concentration
Only five groundwater treatment plants included in the inventory had dissolved oxygen < 1 mg/L. None
of these plants achieved effective manganese removal (Fig. 2.11).
26
Mn removal (%)
100
80
60
NH4+ removal < 85%
pH < 7.1
40
20
0
0
2
4
6
8
10
12
14
Oxygen in filtrate (mg/l)
Figure 2.11: Manganese removal efficiency as a function of dissolved oxygen (mg/L) in filtrate.
It is known from the literature that manganese removal is negatively affected by the absence of, or low
concentration of, i.e., dissolved oxygen (Stumm and Morgan, 1996). Lack of oxygen creates a reducing
environment, Mn2+ cannot be oxidized, and already formed manganese oxides will start leaching Mn2+.
Consequently, a minimal concentration of dissolved oxygen in water is required to prevent manganese oxide
reduction, and to support manganese removal by adsorption and oxidation.
Apart from a minimum value of dissolved oxygen concentration required to achieve effective manganese
removal, no other distinct relationship between manganese removal efficiency and the oxygen level in filtrate
was observed from univariate correlation analysis. Operation of GWTPs with high to very high dissolved
oxygen concentrations does not necessarily lead to complete manganese removal. This observation is backed
up by the relatively high loading coefficient in the second principal component obtained by PCA (see
Table 2.1), indicating that oxygen levels only play an indirect role in manganese removal. More detailed
assessment of the data from plants where manganese removal was ineffective with dissolved oxygen in
filtrate >1 mg/L revealed that one or more of earlier identified critical water quality or process parameters
may explain poor manganese removal. In groundwater, manganese is typically present in low concentrations
and exhibits low oxygen demand (0.29 mg O2 per mg Mn). However, although the oxygen requirement for
oxidation of dissolved manganese is very limited, groundwater entering the filters should have a high
dissolved oxygen concentration to enable oxidation of dissolved iron (0.14 mg O2 per mg Fe2+), NH4+ (3.6
mg O2 per mg of NH4+), and in some cases methane (4.0 mg O2 per mg CH4) and hydrogen sulfide (1.9 mg
O2 per mg H2S), since these compounds are typically oxidized prior to manganese, if not removed during
the aeration (CH4 and H2S).
2.4.9 Effect of filtration type (gravity or pressure)
The type of manganese removal filters applied (gravity or pressure filters) was not included in the statistical
analysis. Additional assessment of available data revealed that the type of filters applied, could have an
important impact on the manganese removal efficiency. In Figure 2.12 (A to D), the effect of some selected
water quality and process parameters on manganese removal efficiency is given, taking into account the type
of filters applied.
27
10
20
30 40 50 60 70 80
NH4 removal efficiency (%)
90
4.5
5.0
5.5
6.0
6.5
pH of filtrate
7.0
7.5
100
90
80
70
60
50
40
30
20
10
0
B 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Fe loading per filter run (kg Fe/m2)
100
90
80
70
60
50
40
30
20
10
0
100
100
90
80
70
60
50
40
30
20
10
0
C 4.0
Mn removal efficiency (%)
100
90
80
70
60
50
40
30
20
10
0
0
A
8.0
D
0
2
4
6
8
10 12 14
Filtration rate (m/h)
16
18
20
Figure 2.12: Mn removal efficiency achieved in gravity (●) and pressure filters (□) in conventional
aeration-filtration GWTPs as a function of A: NH 4 + removal efficiency. B: Fe loading per filter run
(kg Fe/m 2 ). C: pH of filtrate. D: filtration rate (m/h).
The results of the inventory show that a large number of GWTPs utilizing gravity filters achieved complete
manganese removal, while none of the plants having pressure filters achieved that. It should be noted that
pressure filters typically operate at a significantly higher filtration rate (Fig. 2.12D), which results in higher
iron loading and shorter contact time (Fig. 2.12B). However, also pressure filters that were operated at a
filtration rate < 10.5 m/h, did not achieve complete manganese removal possibly due to a somewhat lower
pH (Fig. 2.12C) that could be caused by pressure aeration and (related) lack of CO2 degassing.
2.5 Conclusions
The manganese removal efficiency of (the first stage of) aeration-rapid sand filtration in GWTPs was
assessed. The results from the conducted statistical analysis and assessment of data collected confirmed that
manganese removal in the conventional aeration-rapid sand filtration treatment plants is complex and
simultaneously influenced by several water quality and process parameters. Close inspection of the collected
data from a large number of groundwater treatment plants and conducted statistical analysis (univariate
correlations and PCA) shows that manganese removal efficiency is influenced by several parameters
simultaneously, including both water quality matrix and several (operational) design parameters. The results
of the PCA show that at least four components are necessary to explain at least 75% of the total variance
in the data. Furthermore, the PCA reveals that iron loading, NH4+ removal efficiency and pH in filtrate play
a major role in manganese removal, while oxygen in the filtrate, and phosphate, manganese and NH4+
concentrations in raw water were found to influence manganese removal to a smaller degree.
Univariate statistics and assessment of available data indicate that very effective manganese removal
efficiency in the first aeration-filtration stage can be achieved under the following conditions:
28
pH of filtrate
filtration rate
oxygen in filtrate
:
:
:
:
:
:
> 85%
< 2.7 kg Fe/m2
> 7.1
< 10.5 m/h
> 11.5 min
≥ 1 mg/L
The results of this study might be a helpful tool for engineers to design a traditional aeration-filtration
treatment scheme for proper manganese removal. In addition results emerging from univariate statistics and
assessment of available data from full scale plants could be of help in making a decision if manganese
removal could be combined with removal of iron and ammonium in a single filtration step.
2.6 Acknowledgements
This research was financially and technically supported by WLN and the Dutch water companies
Waterbedrijf Groningen and Waterleiding Maatschappij Drenthe. A large number of Dutch water supply
companies Brabant Water, Dunea, Evides, Oasen, Vitens and WML made this study possible by sharing
data from their plants. We are also grateful to the Belgian water companies PIDPA and VMW, and the
German water companies TAV, SWE, WVO and WAZ Niedergrafschaft, Water supply company of Novi
Sad, Serbia and Water Authority of Jordan.
2.7 References
Bartlett, M. S. 1937. "Properties of sufficiency and statistical tests". Proceedings of the Royal Statistical
Society, Series A 160, 268–282.
Buamah, R., Petrusevski, B., Schippers, J.C. 2008. Adsorptive removal of manganese (II) from the aqueous
phase using iron oxide coated sand. Journal of Water supply: Research and Technology-AQUA 57.1,
1-11.
Buamah, R. 2009. Adsorptive removal of manganese, arsenic and iron from groundwater. PhD thesis,
Wageningen University and UNESCO-IHE, The Netherlands.
Buamah, R., Petrusevski, B., de Ridder, D., van de Watering, S. and Schippers, J.C. 2009. Manganese removal
Technology: Water Supply 9.1, 89 - 98.
Burger, M.S., Krentz C.A., Nercer S.S., Gagnon G.A. 2008. Manganese removal and occurrence of
Cerny, C.A., & Kaiser, H.F. 1977. A study of a measure of sampling adequacy for factor-analytic correlation
matrices. Multivariate Behavioral Research, 12 (1), 43-47.
29
Flemming, H. C., Steele, H., Rott, U., Meyer C. 2004. Optimirung der in-situ reaktortechnologie zur
dezentralen trinkwassergewinnung und grundwasseraufbereitung durch modelhafte untersuchungen
beteiligter biofilme. Report by the Institute for Sanitary Engineering, Water Quality and Solid Waste
Management of the University of Stuttgart.
Gouzinis, A., Kosmidis, N., Vayenas, D.V., Lyberatos, G. 1998. Removal of manganese and simultaneous
removal of NH3, Fe and Mn from potable water using trickling filter. Wat. Res. Vol. 32 (8), 24422450.
Graveland, A. 1971. Removal of manganese from groundwater, PhD thesis, Technical University Delft, The
Netherlands.
Graveland, A., Heertjes, P.M. 1975. Removal of manganese from groundwater by heterogeneous
autocatalytic oxidation, Trans. Instn. Chem. Engrs., 53, 154-164.
Hastings, D., Emerson, S. 1986. Oxidation of manganese by marine bacillus: Kinetic and thermodynamic
considerations. Geochimica et Cosmochimica Acta 50, 1819-1824.
Huysman, K. 2010. Provinciale en Intercommunale Drinkwatermaatschappij der Provincie Antwerpen
(PIDPA), personal communication, Belgium.
Katsoyiannis, I. A., Zouboulis A.I. 2004. Biological treatment of Mn (II) and Fe (II) containing groundwater:
kinetic considerations and product characterization. Water research 38, 1922-1932.
Klewicki, J.K., Morgan J.J. 1999. Dissolution of βMnOOH particles by ligands: pyrophosphate,
ethylenediaminetetraacetate and citrate. Geochim. Cosmochim. Acta, 63, 3017-3024.
O’Melia, C.R., Crapps, D.K. 1964. Some chemical aspects of rapid sand filtration. Journal of American
Water Works Association, October 1964, 1326-1344.
Hu P-Y, Yung-Hsu Hsieh, Jen-Ching Chen, Chen-Yu Chang. 2004a. Adsorption of divalent manganese ion
on manganese coated sand, Journal of Water supply: Research and Technology - AQUA 53.3, 151158.
Hu, Y-H., Hsieh, Y-H., Chen, J-C., Chang, C-Y. 2004b. Characteristics of manganese-coated sand using
SEM and EDAX analysis. Journal of Colloid and interface science, 272, 308-313.
Post, J. E. 1999. Manganese oxide minerals: Crystal structures and economic and environmental significance.
Ramstedt, M., Schukarev, A., Sjoberg, S. 2002. Characterization of hydrous Manganite (γMnOOH) surfaces
– an XPS study. Surf. Interface Anal. 34, 632-636.
Scherer and Wichmann 2000. Treatment of groundwater containing methane – combination of the
processing stages desorption and filtration. Acta Hydrochemica et hydrobiologica, 28 (3), 145-154.
Stumm, W. and Morgan, J.J. 1996. Aquatic Chemistry, chemical equilibria and rates, 3rd ed. Wiley, New
York.
30
UNEP 2000. Vital water graphics, http://www.unep.org/vital water
Vandenabeele, J., Van de Woestyne, M., Houwen, F., Germonpré, R., Vandesande, D., Verstreate, W. 1995.
Role of Autotrophic Nitrifiers in Biological Manganese Removal from Groundwater containing
Manganese and Ammonium. Microbial Ecology 28, 83-98.
Wolthoorn, A., Temminghoff, E.J.M., Weng, L., Riemsdijk, W.H. van 2004. Colloid formation in
groundwater: effect of phosphate, manganese, silicate and dissolved organic matter on the dynamic
heterogeneous oxidation of ferrous iron. Applied Geochemistry 19, 611-622.
31
Figure: Scanning Electron Microscopy (SEM) picture of Birnessite in filter media coating obtained
from the pilot research at Grobbendonk - August 2013
(photo: Arie Zwijnenburg, Wetsus)
32
3
CHARACTERIZATION OF FILTER MEDIA COATING
Jantinus H. Bruins, Branislav Petrusevski, Yness M. Slokar, Joop C. Kruithof, Maria D. Kennedy
(2015). Manganese removal from groundwater: characterization of filter media coating. Desalination &
Water Treatment, 55 (7), 1851 – 1863.
33
3.1 Abstract
Removal of manganese in conventional aeration-filtration groundwater treatment plants (GWTPs) results
in the formation of a manganese oxide coating on filter media. The formation of this coating is an essential
prerequisite for efficient manganese removal. Different manganese oxides have varying affinities for
autocatalytic adsorption/oxidation of dissolved manganese. The aim of this study was to characterize
manganese oxide(s) on filter media from successfully operating manganese removal plants. Characterization
of filter media samples from full-scale groundwater treatment plants and identification of manganese species
was carried out by X-ray diffraction, scanning electron microscopy coupled with energy dispersive Xradiation (SEM-EDX), Raman spectroscopy and electron paramagnetic resonance (EPR). The results
showed that the manganese oxide present in the aged coating was poorly crystalline. Results from the Raman
spectroscopy and the detailed EPR analysis show that the manganese oxide in the ripened coating was of a
Birnessite type, and of physicochemical origin. The results transpiring from this research suggest that the
presence of Birnessite in the coating is essential for effective manganese removal in conventional aerationfiltration treatment plants, since Birnessite has a considerable ability to adsorb and oxidize dissolved
manganese.
Keywords: Groundwater treatment, Manganese removal, MOCS, Filter media characterization,
Birnessite
3.2 Introduction
In European countries the removal of manganese from groundwater is commonly achieved by aerationrapid sand filtration, eliminating the need for strong oxidants to enhance manganese oxidation. This type of
manganese removal is effective and beneficial for both environmental and economic reasons, but requires
a long ripening period of virgin filter media. The ripening time can last from several weeks to more than a
year, before effective manganese removal is achieved (Cools, 2010; Huysman, 2010; Krull, 2010).
Although extensive research has been carried out on manganese removal by aeration-rapid sand filtration,
the mechanisms controlling the ripening period, including the formation of a manganese oxide coating on
virgin filter media, are still not well understood. Several researchers have suggested that the use of
(pre)coated or (‘bio’) aged filter media can shorten the ripening period (Buamah et al., 2008; Buamah, 2009,
Buamah et al., 2009; Hu et al., 2004a; Islam et al., 2010; Kim et al., 2009; Olanczuk-Neyman et al., 2000; Sahabi
et al., 2009; Stembal et al., 2005; Tekerlekopoulou et al., 2005; Tiwari et al., 2007). Aged filter media in
manganese removal filters typically consists of Manganese Oxide-Coated Sand (MOCS) and/or Manganese
Oxide-Coated Anthracite (MOCA). Autocatalytic properties that are attributed to the coating of these filter
media enhance the adsorption of dissolved manganese and its subsequent oxidation. The ripening time of
filter media is controlled by the type and amount of manganese oxide(s) present in the coating. As proposed
by Stumm and Morgan (1996), oxidation of manganese in homogeneous aqueous solution follows different
pathways, as shown in a simplified scheme (Fig. 3.1).
34
Figure 3.1: Simplified scheme of Mn 2 + oxidation pathways according to Stumm and Morgan (1996).
The propensity of manganese present in filter media coating heavily depends on the valence of manganese
in the oxide (Anschutz et al., 2005). A higher valence restricts the possibility of further oxidation. The
valences of manganese oxides shown in Fig. 3.1 are given in Table 3.1.
Table 3.1: Types of manganese oxide and the average valence of manganese in these oxides
Manganese oxide
Pyrochroite
Hausmannite
Manganite / Feitknechtite
Buserite / Birnessite
Nsutite
Pyrolusite
Valence of manganese
2
(Stumm and Morgan)
~ 2.7
(Anschutz et al., 2005)
3
3.5 to 3.9 (Cui et al. 2009)]
~ 4
(Anschutz et al.,,2005)
4
In nature, the most stable form of manganese oxide is Pyrolusite (βMnO2). Because the valence of
manganese in this oxide is 4, no further oxidation can take place. Pyrolusite has a large adsorption capacity,
but no autocatalytic oxidative properties (Buamah, 2009). Consequently, when the adsorption capacity is
exhausted, the removal of manganese stops.
Pyrolusite is formed when (powerful) oxidants such as chlorine (Cl2), chlorine dioxide (ClO2), ozone (O3)
or potassium permanganate (KMnO4) are used. These oxidants are commonly applied in countries such as
the USA (Carlson et al., 1997; Knocke et al., 1991). Pre-oxidation with powerful oxidants can achieve a very
effective manganese removal. However, this type of process is associated with disadvantages such as
continuous oxidant dosage to achieve manganese removal. This process is also associated with high costs
and environmental risks, and requires accurate oxidant dosage: underdosing causes incomplete manganese
oxidation, and consequently a poor manganese removal, while overdosing with permanganate gives the
water a pinkish colour. Therefore in Western European countries, such as The Netherlands, manganese
removal by conventional groundwater treatment with aeration-filtration is preferred.
Based on thermodynamic considerations, the following pathway for manganese oxidation in water is
proposed: (Pyrochroite) – Hausmannite – Manganite – Pyrolusite (Graceland, 1971; Graveland and
Heertjes, 1975). However, based on the pε / pH diagram shown in Fig. 3.2 (Stumm and Morgan, 1996),
chemical formation of Pyrochroite (Mn(OH)2) under common groundwater conditions (pH 6 - 8) is
unlikely. Without a catalyst, the pH must be at least 8.6 to achieve Hausmannite (Mn3O4) formation, which
is subsequently removed by filtration (Graveland, 1971). Formation of Pyrochroite requires an even higher
pH, of at least 11.
35
Figure 3.2: Diagram of electron activity (pε) and redox potential (Eh in V) as a function of the pH
showing the stability zones of manganese-containing compounds in aqueous solution (adopted from
Stumm and Morgan, 1996).
The exact pathway by which manganese oxidation occurs under conditions commonly applied in
groundwater treatment, is not known. Besides chemical oxidation, ‘biology’ may play an important role.
Identification of the manganese oxide(s) present in filter media coatings from manganese removal filters
might give insight into the formation of manganese oxide coatings on virgin filter media. This may elucidate
the mechanisms controlling the start-up of the manganese removal process (Hu et al., 2004b). Based on
these results, recommendations for process modifications to shorten the ripening period of manganese
removal filters may be made.
The aim of this study was to characterize the coating of MOCS and MOCA from (ripened) manganese
removal filters in full-scale groundwater treatment plants (GWTPs), by identification of the types of
manganese oxide present in the coating. It was hypothesized that identification of the manganese oxide
present in ripened media from manganese removal filters may help to better understand the driving force
behind efficient manganese removal in conventional aeration-filtration groundwater treatment plants.
MOCS and MOCA characterized in this study were collected from two GWTPs of two water companies as
follows.
•
MOCS from GWTP De Punt (Groningen, The Netherlands).
The sample was taken from a depth of 140-150 cm, from a filter bed that was removed after an
operating time of approximately 15 years.
•
MOCA from GWTP Grobbendonk (Pidpa, Belgium).
The sample was taken from the top of a filter bed, in operation for more than 12 years.
36
Both GWTPs utilize conventional groundwater treatment by aeration-filtration and achieve complete
manganese removal. Information on feed water quality, process design parameters and operational
conditions of full-scale plants De Punt and Grobbendonk are shown in Table 3.2.
Table 3.2: Feed water quality, process design parameters and operational conditions of full scale
GWTPs De Punt and GWTP Grobbendonk
Parameters feed water /
technical information filter
Iron
Manganese
Ammonium
pH
Oxygen
Redox potential
Type of aeration
Position of filter
Type of filtration
Type of filter media
Grain size fraction virgin
media
Filter area
Filter bed height
Flow per filter
Filtration rate
Empty bed contact time
Backwash (BW) criterion
Backwash frequency
Filter bed expansion during
BW
Filtered volume between BW
Iron loading per filter run
(FR)
Unit
De Punt
Grobbendonk
mg/L
mg/L
mg/L
[-]
mg/L
mV
mm
4.5 - 6.9
0.18 - 0.25
0.29 - 0.78
7.3 -7.5
8 - 10
-50 to +50
spray
pre-filter
down flow
quartz sand
1.8 - 2.4
0.03 - 0.14
0.12 - 0.18
< 0.05 - 0.23
7.5 - 7.6
> 10
+ 200 to +300
cascade
post-filter
down flow
anthracite / quartz sand
0.8-1.8 / 0.4-0.8
m2
m
m3/h
m3/m2.h
min
n/week
-
12.5
2
60
4.8
25
head loss
2
no
37.5
1.1 (0.6+0.5)
190
5.0
13.2
head loss
0.5
yes (anthracite)
m3 per filter run
kg Fe/m2.FR
5,000 – 7,000
2.5
10,000
< 0.1
To characterize the filter media coatings and to identify the manganese oxide(s), the following
supplementary techniques were used:
•
•
•
•
Raman spectroscopy;
X-ray Diffraction (XRD);
Scanning Electron Microscopy coupled with Energy Dispersive X-radiation analysis (SEM-EDX)
and
Electron Paramagnetic Resonance (EPR).
As these techniques are supplementary, for complete identification and determination of the origin of
produced MnOx, application of all these methods is required. Use of one single method is insufficient to
characterize and identify the manganese oxides.
37
For additional information with respect to Raman spectroscopy (selection spectroscopy settings - Section
3.1) and SEM-EDX (measuring trace (counter) elements) Section 3.3), additional MOCS samples from post
filters of two other plants were used, as follows.
•
MOCS from GWTP Onnen (Groningen, The Netherlands).
The sample is taken from the top of the filter, in operation for more than 40 years.
•
MOCS from GWTP Wierden (Vitens, The Netherlands).
The sample is taken from the top of the filter, in operation for more than 18 years.
3.3.1 Raman spectroscopy
With Raman spectroscopy it is possible to distinguish different (general) types of manganese oxides. For
the Raman spectroscopy analysis a Horiba Yobin Yvon Labram instrument was used with the following
settings (Table 3.3).
Table 3.3: Parameter settings for Raman spectroscopy analysis.
Parameter
Exposure
Current
Confocal hole
Slit
Laser wave length
Grating
Objective
Density filter
Detector
Detector size
Setting
≥ 30 sec
0.05 mA
1,000 µm
100 µm
532.13 nm
600
x50
D3
synapse CCD
1,024 pixels
Exposure of the samples to high power laser radiation (5 mA) during Raman analysis for a long time (> 120
seconds) may change the structure of the manganese oxides. However, under the applied conditions
(Table 3.3), no structural damage was detected during the tests. Detailed information on the effect of high
power radiation on structural changes of the manganese oxides is given in Section 3.4.1. Before analysis, the
samples (either the integral grain or powder coating) were dried at room temperature to avoid excessive
fluorescence caused by the presence of water in the samples.
3.3.2 XRD
XRD, can be used to determine whether manganese oxide is crystalline or amorphous. Furthermore, this
technique can provide additional information to clarify the sharpness of the peaks found with Raman
spectroscopy. Most XRD measurements were carried out with a Bruker-D8 Advance diffractometer in
Bragg-Brentano focusing geometry. The instrument was equipped with a Vantec PSD detector.
Measurements were carried out at room temperature with the use of monochromatic CoKα1 radiation (λ =
0.179026 nm). The 2θ scan was made in the range 10-110°2θ, using a step size of 0.038°2θ. To check some
of the analyses and to enhance their performance, some XRD measurements were also carried out with a
Bruker-AXS D5005 diffractometer in Bragg-Brentano focusing geometry, equipped with a graphite
monochromator in the diffracted beam. Before analysis, the samples were dried at room temperature and
then pulverized to a size of 10-50 µm. The sample powder (for MOCS only the coating and for MOCA the
38
coating including the anthracite grain) was put in an aluminum sample holder. The 2θ scan was made in the
range 10-110°2θ, using a step size of 0.025°2θ and a counting time of 2 s per step. The radiation was CuKα1
(0.15406 nm). Results obtained by both the XRD instruments were evaluated with the internal Bruker EVA
software.
3.3.3 SEM-EDX
SEM pictures show the structure of the coating and the manganese oxide inside. In addition EDX provides
information about the trace (counter) elements present in these structures. Therefore, differences in peak
ratio found with Raman spectroscopy can be clarified. The SEM pictures were made with a JEOL-6480LV.
Samples (either the integral grain or powder coating) were placed on 15 mm diameter mounts with a doublesided carbon adhesive tab. The SEM was operated both under High Vacuum (HV) and Low Vacuum (LV)
conditions.
Before observation under HV conditions, the samples were coated in a JEOL JFC-1200 fine coater with a
thin (10 nm) Au layer. The operation under HV conditions was done at 6 kV at 10 mm Working Distance
(WD) and Spot Size (SS) 20. Samples were also investigated without Au coating under LV conditions at 10
kV and SS 60. Composition of the adsorbed coating layers was determined with an X-ray microanalysis
(EDX) System type Noran System SIX from the Thermo Electron Corporation. Analyse were done at 10
kV and SS 70 for both the Au-coated samples at HV conditions and non-coated samples at LV conditions.
In the latter, no Au peaks were observed in the EDX spectrum.
3.3.4 EPR
Once the MnOx is characterized as ‘a type of Birnessite’, EPR can be employed to identify the exact type of
Birnessite and additionally also to determine the origin of the Birnessite. For EPR the whole grain was
always used.
EPR analyses were carried out with a Bruker instrument operating at 9.46 GHz (W-waveband). The analyses
were carried out at two different temperatures, i.e., at 77 K (cooled with nitrogen) and 298 K (room
temperature).
3.4.1 Selection Raman spectroscopy settings
As mentioned in Section 3.3.1, the structure of almost all manganese (oxide) compounds (except
Hausmannite) is subjected to Raman spectroscopic changes when exposed to a high current (Julien et al.,
2003, 2004). Laser irradiation in excess of 5 mA for longer than 120 s causes structural changes of the
exposed compounds (polymorphism). To prevent these changes the current of the laser in this study was
limited to 0.05 mA. The differences in spectra obtained with high (5 mA, >120 s) and low (0.05 mA)
radiation intensities are shown in Fig. 3.3.
39
200
400
600
800
1000
1200
Raman shift (cm-1)
low radiation
high radiation
Figure 3.3: Raman spectra at 532 nm of MOCS (Onnen), exposed to low (light blue line) and high
laser radiation (red line).
From the two spectra in Fig. 3.3 it can be seen that high intensity laser irradiation changes the position of
the peaks, as well as the pattern of the spectrum. According to (Julien et al., 2003, 2004), high current
irradiation transforms all MnOx samples (and therefore also Birnessite) into Mn3O4 (Hausmannite).
Hausmannite is very stable when exposed to a Raman laser, so its structure is not changed by the high level
of irradiation. Furthermore, due to its crystalline lattice structure it displays a sharp peak in the Raman
spectrum. To ascertain this transformation, Raman profiles of Birnessite from this study and an MnO2
reference sample (Alfa Aesar) were exposed to high current radiation (5 mA, >120 s), and compared to a
Hausmannite spectrum from the RRUFF database (Downs, 2006) (Fig. 3.4).
200
400
600
800
1000
1200
Raman shift (cm-1)
Figure 3.4: Raman spectra (at 532 nm) of Birnessite from a sample of this study (red line) and
MnO 2 reference (green line) after exposure to 5 mA for >120 s, compared to a Hausmannite spectrum
(purple line) from the RRUFF database (Downs, 2006).
40
From Fig. 3.4 it can be seen that the spectra of Birnessite in a sample from this study as well as the MnO2
reference, exposed to high laser radiation, show a close resemblance to the spectrum of Hausmannite. This
indicates that both manganese oxides (Birnessite and MnO2) underwent structural changes when exposed
to high laser power, as stated by (Julien et al., 2004).
3.4.2 Characterization of MOCA and MOCS by Raman spectroscopy
200
400
600
800
1000
1200
Raman shift (cm-1)
MOCS De Punt
MOCS Wierden
MOCA Grobbendonk
MOCS Onnen
Figure 3.5: Raman spectra (at 532 nm) of the four MOCS / MOCA samples .
The Raman spectra of manganese-coated samples from four GWTPs are presented in Fig. 3.5. From
Fig. 3.5, it can be seen that the spectral patterns of all samples were similar, indicating a similar type of
manganese oxide. The three most pronounced absorbance peaks were found at Raman shifts of 495-505
cm-1, 570-575 cm-1, and 635-655 cm-1, typical for the Birnessite group of manganese oxides (Julien, et al.,
2003). All spectra exhibited a less pronounced peak at a Raman shift at 400-420 cm-1. The undulating peaks
were typical for oxides with a poorly crystalline structure, such as Birnessite (Post, 1999).
To confirm the presence of Birnessite, the spectra of the MOCS/MOCA samples were compared with the
spectra of synthetically produced Birnessite (Ma et al., 2007), shown in Fig. 3.6A. Comparison of the spectra
produced in this study with previously reported results strongly suggested that the manganese oxide in the
coating of the MOCS and MOCA samples was of a Birnessite type.
The spectra of MOCS and MOCA samples were also compared with a Raman profile of a reference sample
of naturally formed Birnessite (Fig. 3.6B), obtained from the Aufgeklärt Glück mine in Hasserode, the Harz
Mountains, Sachsen-Anhalt, Germany (Witzke, 2012).
41
B
200 300 400 500 600 700 800 900
Raman shift (cm-1)
Figure 3.6: A: Raman spectra of synthetically produced Birnessite at 514 nm (Ma et al., 2007). B:
Raman spectra of samples from this study (green, red, blue and yellow lines) compared to the spectrum of
the Birnessite reference (black line).
Fig. 3.6B shows that the main peaks from the Birnessite reference as well as the MOCA and MOCS samples
were on the same Raman shift. Only the ratio and height of the peaks varied between the different samples
and the reference, probably due to a different concentration of counter ions in the materials (Julien et al.,
2003, 2004), as shown by SEM-EDX (Section 3.4.4, Table 3.4).
The Raman spectra strongly suggested that the manganese oxides in the MOCA and MOCS samples were
of a Birnessite type.
In order to exclude the presence of other manganese oxides in the coatings, the Raman spectra of samples
from this study were compared with the reference spectra of four other manganese oxides Pyrochroite (A),
Hausmannite (B), Manganite (C) and Pyrolusite (D) (Fig. 3.7). Spectra A-C were taken from the RRUFF
database (Downs, 2006) and spectrum D from the measured reference sample (Alfa Aesar).
A
200
400
MOCS
600
800
Raman shift (cm-1)
MOCA
1000
Pyrochroite (RUFF)
1200
B
200
400
MOCS
600
800
Raman shift (cm-1)
MOCA
1000
1200
Hausmannite (RUFF)
42
C
200
400
600
800
Raman shift (cm-1)
MOCS
MOCA
1000
1200
D
200
Manganite (RUFF)
400
600
800
Raman shift (cm-1)
MOCS
MOCA
1000
1200
Pyrolusite (Alfa Aeser)
Figure 3.7: Raman spectra at low power and 532 nm of MOCS De Punt and MOCA Grobbendonk,
compared to (A) Pyrochroite, (B) Hausmannite, (C) Manganite and (D) Pyrolusite.
The four manganese oxides displayed one peak in the same region as the MOCS/MOCA samples between
630 and 650 cm-1. This wavelength is characteristic for all manganese oxides and not indicative of a particular
one. Therefore, it was concluded that apart from Birnessite no other manganese oxides were present in the
coating of the filter media.
3.4.3 XRD
Fig. 3.8 shows an XRD spectrum of MOCS coating from GWTP De Punt and a MOCA coating from
GWTP Grobbendonk. The analyzed MOCA sample contained some crushed anthracite (carbon) since it
was impossible to separate the coating completely from the anthracite.
20
30
40
50
60
70
80
2-theta
coating of MOCS (De Punt)
MOCA (Grobbendonk)
Figure 3.8: XRD spectra (conducted with a Bruker D8), of pulverized MOCS coating and MOCA
(the latter including crushed anthracite core).
From Fig. 3.8 it can be seen, as already shown by the Raman spectroscopy, that the MOCS coating was
poorly crystalline. The only peak (31⁰, 2θ) belonged to silica originating from the filter media. Additional
(sharp) peaks were not present in the spectra from the MOCS sample, so no crystalline MnOx was present
on the MOCS. Also no crystalline MnOx was found on the MOCA. The few small peaks and the broad
peak in the MOCA sample (25⁰, 2θ) originate from carbon or graphite (C) from crushed anthracite that
could not be removed completely from the coating (as mentioned above).
43
For comparison, the XRD spectra of three Alfa Aesar reference manganese oxides i.e., MnO2, Mn2O3, and
MnO with crystalline structure are presented in Fig. 3.9.
20
30
40
50
60
70
80
90
100
2-theta
MnO
Mn2O3
Mn2O3
MnO2
MnO2
Figure 3.9: XRD spectra of reference manganese oxides: MnO, Mn 2 O 3 and MnO 2 (Alfa Aesar).
Contrary to the MOCA and MOCS coating, the XRD spectra of the three reference manganese oxides
showed sharp peaks and the 2θ positions of the peaks were confirmed by the reference XRD spectra of
oxides from the internal Bruker EVA database. If the manganese oxide in the samples from this study had
been crystalline, sharp peaks should have been observed similar to the spectra of the three reference
manganese oxides. The absence of sharp peaks confirmed the amorphous character of the sample coating.
Thus the XRD results also confirmed that the manganese oxide(s) in the MOCS and MOCA samples were
not crystalline manganese oxides (i.e., MnO2, Mn2O3 or MnO).
3.4.4 SEM-EDX
Figure 3.10: SEM image of virgin quartz media (sand); 500x magnification.
In Fig. 3.10 a SEM image of virgin quartz (sand) filter media is shown. Fig. 3.10 shows that the surface of
virgin sand was not completely smooth, indicating a high porosity and specific surface area, and therefore
probably contained many sites for attachment of bacteria and/or manganese, which could shorten the
ripening time.
44
In Fig. 3.11 SEM images of MOCS and MOCA samples are shown.
A
B
Figure 3.11: SEM images of filter media coating. A - MOCS De Punt; 10,000x.
B - MOCA Grobbendonk; 5,000x.
Images A and B in Fig. 3.11 confirm the Raman spectroscopy results and the XRD analysis that the
manganese coating was poorly crystalline.
Results of the SEM-EDX analyses show that apart from manganese and iron, the coating of the filter media
contained (trace) elements such as aluminum, calcium, magnesium, potassium, sodium and/or silica
(Table 3.4).
Table 3.4: SEM-EDX analysis of counter (trace) elements in weight (%).
Element
Aluminum (Al)
Calcium (Ca)
Magnesium (Mg)
Potassium (K)
Silica (Si)
Sodium (Na)
MOCA
Grobbendonk
0.5
6.8
0.3
0.2
1.0
< 0.1
MOCS De Punt
< 0.1
2.3
< 0.1
< 0.1
5.9
< 0.1
MOCS Onnen
0.6
7.2
0.4
< 0.1
2.6
0.2
MOCS Wierden
< 0.1
7.7
0.3
< 0.1
0.9
< 0.1
The counter (trace) elements in Table 3.4 are of importance with respect to the ratio and height of the
Raman peaks, as discussed in Section 3.4.2 and shown in Fig. 6B. Counter ions present in filter media
coatings may also impact the adsorptive properties of the coating.
3.4.5 EPR
In Fig. 3.12 the EPR spectra of MOCS De Punt measured at two temperatures (298 K - room temperature
and 77 K) are shown. Due to the analytical interference of carbon (originating from anthracite) it was not
possible to obtain a complete EPR spectrum for MOCA Grobbendonk.
45
150
298 K
g = 1.956
ΔH = 2642 Gauss
100
g at 298 K calculated at H0 (3455 Gauss)
50
0
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
5500
6000
-50
-100
77 K
-150
298 K
150
100
50
0
3100
3150
3200
3250
3300
3350
3400
3450
3500
3550
3600
3650
3700
3750
3800
-50
-100
-150
Figure 3.12: EPR spectra of MOCS De Punt. Top half: spectra measured at 298 K (blue line) and
77 K (green line). Bottom half: expanded spectrum measured at 77 K.
The detailed EPR pattern (Fig. 3.12, the bottom half) shows the ‘6 line hyperfine structure of manganese’.
This pattern is typical for oxides containing manganese ions with a different valence, including Mn2+
(Browne, 2011). Birnessite has a valence between 3.5 and 3.9 (Cui et al., 2009; Post, 1999; Kim et al., 2011),
due to the presence of Mn3+ and Mn4+ in the lattice. This pattern is an additional indication that the
predominant manganese component of the MOCS coating is a manganese oxide, containing manganese
ions with different oxidation numbers, such as Birnessite.
Generally, EPR spectra are characterized by two parameters: (i) the line width: ΔH in gauss, and (ii) the
spectroscopic splitting factor: g (g-factor). EPR was used to investigate the origin of different types of
Birnessite (Kim et al., 2011). These values of ΔH measured at 77 K and 298 K and the g-factor calculated at
298 K are characteristic of Birnessite.
ΔH is measured as the distance between the highest and lowest point of the signal wave, and the g factor is
calculated from the spectrometer microwave frequency (y) and the magnetic field (H0) applied during the
analysis when the signal is zero (Eq. 1):
h
y
×
g=
(1)
μB H 0
where:
h
µB
y
H0
=
=
=
=
Planck constant (6.62×10−34 Js);
Bohr magneton (9.27×10−28 J/G);
spectrometer microwave frequency (GHz);
applied magnetic field where signal is zero (gauss).
46
ΔH values measured at 298 K and at 77 K were 2,642 gauss and 3,950 gauss, respectively (Fig. 3.12 - top
half). Taking into account the applied magnetic field (H0) of 3,455 gauss at 298 K, the calculated g factor
was 1.956. Comparing the EPR results from this study with the data reported by (Kim et al., 2011), a few
conclusions can be drawn. Firstly, the EPR analysis confirmed the results obtained by Raman spectroscopy
that the manganese oxide on MOCS is Birnessite. Next, comparing the two ΔH values and the g factor with
the results from Kim et al. (2011) shows that the Birnessite in the coating of MOCS De Punt was of
physicochemical origin. However, this does not mean that the growth of manganese oxide on the virgin
filter media (ripening process) starts exclusively chemically. The MOCS sample from GWTP De Punt was
taken after more than 15 years of operation, and consequently the EPR analysis only confirmed the
physicochemical nature of Birnessite after this elapsed time.
3.4.6 The importance of Birnessite formation on MOCA/MOCS for
manganese removal
The mineral Birnessite was first found in 1956 in a small hamlet called Birness, Scotland, UK (Jones and
Milne, 1956). At that time it was a new mineral to which the name Birnessite, after the locality, was given.
Its molecular structure was given as (Na0.7Ca0.3)Mn7O14 . 2.8H2O. In time, different formulas were proposed
for Birnessite, indicating there is no single structure, but different compositions exist (‘minerals of the
Birnessite type’). From this study it can be observed that Birnessite was the manganese compound formed
during filter ripening, and predominantly present in all manganese oxide coatings extracted from the four
GWTPs that showed complete manganese removal. Consequently, Birnessite appears to be of particular
importance for manganese removal in conventional aeration-filtration groundwater treatment plants. The
presence of Birnessite may provide an explanation for the very effective manganese removal in these filters
in practice. Birnessite is extremely suitable to remove manganese, because of its structure as described by
Post (Post, 1999): ‘the Birnessite group of minerals has layered structures, which may readily undergo
oxidation-reduction and cation-exchange reactions and play a major role in controlling groundwater
chemistry’. The high cation exchange and adsorption capacity of Birnessite is also described by several other
researchers (Kim et al., 2009; Golden et al., 1986; Han et al., 2006; Kim et al., 2008; Lee et al., 2009; Murray et
al., 1976; Pretorius and Linder, 2001; White, 1997). Thus, Birnessite is very suitable to adsorb Mn2+.
The Birnessite group of minerals (including Buserite) is also important, because of its high reactivity (Post,
1999; Post and Veblen, 1999). In Buserite and Birnessite the average valence number of manganese ranges
from +3.5 to +3.9 (Cui et al., 2009). Although restricted compared to other manganese oxides, further
oxidation of Buserite and Birnessite is still possible. The combination of properties makes Birnessite a highly
reactive manganese oxide, with very good adsorptive properties for dissolved manganese, and its subsequent
autocatalytic oxidation. The knowledge that Birnessite is the manganese oxide responsible for efficient
manganese removal can help substantially accelerate the ripening process of virgin filter media, by creating
conditions favouring the formation of Birnessite. Knowing that the oxidation from Mn2+ into Pyrolusite
(Stumm and Morgan, 1996), via Birnessite (Pyrochroite → Buserite / Birnessite → Nsutite → Pyrolusite,
Fig. 3.2), is only possible under (very) alkaline conditions (Feng et al., 2005), it is not likely that a fast filter
ripening of virgin filter media starts in a chemical way, without oxidant dosage. Therefore the formation of
the manganese coating may be initiated by bacterial activity. Several researchers have suggested that
manganese removal is influenced by bacteria (e.g., Pseudomons sp., Leptothrix sp.), which are able to oxidize
Mn2+ (Kim et al., 2011; Barger et al., 2005; Burger et al., 2008a, 2008b; Geszvain, 2011; Katsoyiannis and
Zouboulis, 2004; Tebo et al., 2004; Villalobos et al., 2003; Vandenabeele et al., 1992, 1995; Vandenabeele,
1993, Tekerlekopoulou et al., 2008). Therefore the ripening of the filter media probably starts with the
47
biological formation of Birnessite. In time very fast physicochemical auto-catalytic adsorption/oxidation
reactions may become more important and result in production of Birnessite of physicochemical origin,
whose presence is shown in this study. However, further research is required to support this hypothesis.
3.5 Conclusions
The Raman spectroscopy, XRD and SEM analyses carried out in this study showed that the manganese
oxide in the coating of the manganese removing filter media is poorly crystalline. Raman spectroscopy and
EPR analysis further clarified that the predominant manganese oxide, responsible for effective removal of
dissolved manganese, is of a Birnessite type. Calculation of ΔH and the g factor from EPR analysis and
comparison of these parameters with results from literature identified Birnessite as being of physicochemical
origin, but the sampling after a ripening period of about 15 years does not exclude the possibility of
Birnessite formation starting via a biological pathway. Despite the generally accepted theory that the
manganese oxidation pathway is via Hausmannite and Manganite, the results transpiring from this research
imply that in water treatment practice oxidation of manganese on the surface of manganese removal filter
media is more likely to form a Birnessite type of manganese oxide. Birnessite has very good properties for
adsorption and autocatalytic oxidation of dissolved manganese. Identification of Birnessite as the
predominant manganese oxide in filter media that effectively remove manganese could possibly enable
shortening ripening time in conventional aeration-filtration groundwater treatment plants by creating
conditions that favour the formation of this compound.
This research is financially and technically supported by WLN and by the Dutch water companies
Waterbedrijf Groningen (WBG) and Waterleiding Maatschappij Drenthe (WMD). The authors of this paper
would like to thank Mr. Arie Zwijnenburg and Mr. Ton van der Zande from Wetsus, Centre of Excellence
for Sustainable Water Technology, Mr. Wesley Browne from the University of Groningen and Mr. Ruud
Hendrikx of the Department of Materials Science and Engineering from the Technical University of Delft
for their help with the analyses of samples and interpretation of the results. The authors would also like to
thank the Belgian water company Pidpa and the Dutch water company Vitens for providing filter media for
this study and their willingness to disclose their groundwater treatment plant data. Finally we would like to
thank Dr. Thomas Witzke (mineralogist) for donating a naturally formed Birnessite reference sample.
3.7 References
Anschutz, P., Dedieu, K., Desmzes, F. and Chaillou, G. 2005. Speciation, oxidation state, and reactivity of
particulate manganese in marine sediments. Chem. Geol., 218, 265-279.
Barger, J.R, B.M. Tebo, B.M., Bergmann, U., Webb, S.M., Glatzel, P., Chiu, V.Q., and Villalobos, M. 2005.
Biotic and abiotic products of Mn (II) oxidation by spores of the marine Bacillus sp. strain SG-1. Am.
Mineral., 90, 143-154.
48
Browne, W.R., 2011. Assistant professor, Stratingh Institute for Chemistry, University of Groningen,
personal communication, The Netherlands.
Buamah, R., Petrusevski B. and Schippers, J.C. 2008. Adsorptive removal of manganese (II) from the
aqueous phase using iron oxide coated sand. J. Water Supply Res. T., 57.1, 1-11.
Buamah, R. 2009. Adsorptive removal of manganese, arsenic and iron from groundwater, PhD thesis,
UNESCO-IHE Delft / Wageningen University, The Netherlands.
Buamah, R., Petrusevski, B., de Ridder, D., van de Wetering, S. and Schippers J.C. 2009. Manganese removal
in groundwater treatment: practice, problems and probable solutions, Water Sci. Technol., 9.1, 8998.
Burger, M.S., Mercer, S.S., Shupe G.D. and Gagnon G.A. 2008a. Manganese removal during bench-scale
biofiltration. Water Res., 42, 4733-4742.
Burger, M.S., Krentz, C.A., Mercer S.S. and Gagnon G.A. 2008b. Manganese removal and occurrence of
manganese oxidizing bacteria in full-scale biofilter. J. Water Supply Res. T., 57.5, 351-359.
Carlson, K.H., Knocke W.R. and Gertig K.R. 1997. Optimizing treatment through Fe and Mn fractionation,
J. Am. Water Works Assoc., 89.4, 162-171.
Cools B. 2010. Vlaamse Maatschappij voor Watervoorziening (VMW), personal communication, Belgium.
Cui, H., Qiu, G., Feng, X., Tan, W. and Liu F. 2009. Birnessite with different average manganese oxidations
states synthesized, characterized, and transformed to Todorokite at atmospheric pressure, Clay. Clay
Miner., 57.6, 715-724.
Downs, R.T. 2006. The RRUFF Project: An integrated study of the chemistry, crystallography, Raman and
infrared spectroscopy of minerals, program and abstracts of the 19th general meeting of the
international mineralogical association in Kobe, Japan.
Feng, X., Tan, W., Liu, F., Huang Q. and Liu, X. 2005. Pathways of Birnessite formation in alkali medium.
Sci. China Ser. D., 48.9, 1438-1451.
Geszvain K., Yamaguchi, Ai., Maybee, J. and Tebo, B.M. 2011. Mn (II) oxidation in Pseudomonas putida GB1 is influenced by flagella synthesis and surface substrate, Arch. Microbiol., 193, 605-614.
Golden, D.C., Dixon, J.B. and Chen, C.C. 1986. Ion exchange, thermal transformations, and oxidizing
properties of Birnessite. Clay. Clay Miner., 34.5, 511-520.
Graveland, A. 1971. Removal of manganese from groundwater, PhD thesis, Technical University Delft, The
Netherlands.
Graveland, A. and Heertjes P.M. 1975. Removal of manganese from groundwater by heterogeneous
autocatalytic oxidation. Trans. Inst. Chem. Eng., 53, 154-164.
49
Han, R., Zou, W., Zhang, Z., Shi, J. and Yang, J. 2006. Removal of copper (II) and lead (II) from aqueous
solution by manganese oxide coated sand. I. Characterization and kinetic study. J. Hazard. Mater,
137, 384-395.
Hu, P.Y., Hsieh, Y.H., Chen J.C. and Chang, C.Y. 2004a. Adsorption of divalent manganese ion on
manganese-coated sand, J. Water Supply Res. T., 53.3, 151-158.
Hu, P.Y., Hsieh, Y.H., Chen J.C. and Chang, C.Y. 2004b. Characteristics of manganese-coated sand using
SEM and EDAX analysis. J. Colloid Interf. Sci., 272, 308-313.
Huysman, K. 2010. Provinciale en Intercommunale Drinkwatermaatschappij der Provincie Antwerpen
(PIDPA), personal communication, Belgium.
Islam, A.A., Goodwill, J.E., Bouchard, R., Tobiasen J.E. and Knocke W.R. 2010. Characterization of filter
media MnO2(s) surfaces and Mn removal capability. J. Am. Water Works Assoc., 102.9, 71-83.
Julien, C., Massot, M., Baddour-Hadjean, R., Franger, S., Bach, S. and Pereira-Ramos, J.P. 2003. Raman
spectra of Birnessite manganese dioxides. Solid State Ionics, 159, 345-356.
Julien, C., Massot, M. and Poinsignon, C. 2004. Lattice vibrations of manganese oxides. Part I. Periodic
structures. Spectrochim. Acta A, 60, 689-700.
Jones, L.H.P. and Milne, A.A. 1956. Birnessite, a new manganese oxide mineral from Aberdeenshire,
Scotland, Mineral. Mag. Journal Min. Soc., XXXI.235, 283-288
Katsoyiannis I.A. and Zouboulis, A.I. 2004. Biological treatment of Mn (II) and Fe (II) containing
groundwater: Kinetic considerations and product characterization. Water Res., 38, 1922-1932.
Kim, S.S., Bargar, J.R., Nealson, K.H., Flood, B.E., Kirschvink, J.L., Raub, T.D., Tebo B.M. and Villalobos
M. 2011. Searching for bio signatures using electron paramagnetic resonance (EPR) analysis of
manganese oxides. Astrobiology, 11.8, 775-786.
Kim, I. and Jung, S. 2008. Soluble manganese removal by porous media filtration. Environ. Technol., 22,
1265-1273.
Kim, W.G., Kim, S.J., Lee, S.M. and Tiwari, D. 2009. Removal characteristics of manganese-coated solid
samples for Mn (II). Desal. Wat. Treat., 4, 218-223.
Knocke, W.R., van Benschoten J.E., Kearney, M.J., Soborski A.W. and Reckhow, D.A. 1991. Kinetics of
manganese and iron oxidation by Potassium Permanganate and chlorine dioxide. J. Am. Water Works
Assoc., 83.6, 80-87.
Krull, J. 2010. Stadtwerke Emden (SWE), personal communication, Germany.
Lee, S.M., Tiwari, D., Choi, K.M., Kim, W.G., Yang, J.K. and Lee, H.D. 2009. Removal of Mn (II) from
aqueous solutions using manganese coated sand samples, J. Chem. Eng. Data, 54, 1823-1828.
Ma, S.B., Ahn, K.Y., Lee, E.S., Oh, K.W. and Kim, K.B. 2007. Synthesis and characterization of manganese
dioxide spontaneously coated on carbon nanotubes. Carbon, 45, 375-382.
50
Murray, J. 1976. The interaction of metal ions at the manganese dioxide-solution interface. Geochim.
Cosmochim. Ac., 39, 606-619.
Olanczuk-Neyman, K. and R. Bray 2000. The role of physico-chemical and biological processes in
manganese and ammonia nitrogen removal from groundwater. Pol. J. Environ. Stud., 9.2, 91-96.
Post, J.E. 1999. Manganese oxide minerals: Crystal structures and economic and environmental significance,
P. Natl. Acad. Sci. USA, 96, 3447-3454.
Post, P.E. and Veblen, D.R. 1999. Crystal structure determinations of synthetic sodium, magnesium, and
potassium Birnessite using TEM and the Rietveld method.Am. Mineral., Volume 7.5, 477-489.
Pretorius, P.J. and Linder P.W. 2001. The adsorption characteristics of δ-Manganese dioxide: A collection
of diffuse double layer constants for the adsorption of H+, Cu2+, Ni2+, Zn2+, Cd2+ and Pb2+. Appl.
Geochem., 16, 1067-1082.
Sahabi, D.M., Takeda, M., Suzuki, I. and Koizumi, J.I. 2009. Removal of Mn2+ from water by ‘aged’ biofilter
media: The role of catalytic oxides layers. J. Bio. Sci. Bioeng., 107.2, 151-157.
Stembal, T., Markic, M., Ribicic, N., Briski, F. and Sipos, L. 2005. Removal of ammonia, iron and manganese
from ground waters of Northern Croatia – pilot plant studies, Process Biochem., 40, 327 -335.
Stumm, W. and Morgan, J.J. 1996. Aquatic chemistry, chemical equilibria and rates, 3rd ed. Wiley, New York,
464-467.
Tebo, B.M., Marger, J.R., Clement, B.G., Dick, G.J., Murray, K.J., Parker, D., Verity, R. and Webb, S.M.
2004. Biogenic Manganese oxides: Properties and mechanisms of formation, Annu. Rev. Earth Pl.
Sci., 32, 287-328.
Tekerlekopoulou, A.G., Vasiliadou I.A. and Vayenas, D.V. 2008. Biological manganese removal from
potable water using trickling filters. Biochem. Eng. J., 38, 292-301.
Tekerlekopoulou A.G. and Vayenas, D.V. 2008. Simultaneous biological removal of ammonia, iron and
manganese from potable water using a trickling filter. Biochem. Eng. J., 39, 215-220.
Tiwari, D., Yu, M.R., Kim, M.N., Lee, S.M., Kwon, O.H., Choi, K.M., Lim G.J.and Yang, J.K. 2007.
Potential application of manganese coated sand in the removal of Mn (II) from aqueous solutions.
Water Sci. Technol., 56:7, 153-160.
Vandenabeele, J., de Beer, D., Germonpré R. and Verstreate, W. 1992. Manganese oxidation by microbial
consortia from sand filters, Microb. Ecol., 24, 91-108.
Vandenabeele, J., van de Woestyne, M., Houwen, F., Germonpré, R., Vandesande D. and Verstreate, W.
1995. Role of autotrophic nitrifiers in biological manganese removal from groundwater containing
manganese and ammonium. Microb. Ecol., 28, 83-98.
Vandenabeele, J. 1993. Manganese removal by microbial consortia from rapid sand filters treating water
containing Mn2+ and NH4+, PhD thesis, Gent, Belgium.
51
Villalobos, M., Toner, B., Barger J. and Sposito, G. 2003. Characterization of the manganese oxide produced
by Pseudomonas putida strain MnB1. Geochim. Cosmochim. Ac., 67(14), 2649-2662.
White D.A. and Asfar-Siddique, A. 1997. Removal of Manganese and iron from drinking water using
hydrous Manganese Dioxide. Solvent Extr. Ion Exc., 15(6), 1133-1145.
Witzke, T. 2012. mineralogist PANanalytical, personal communication, Germany.
52
53
Figure: Scanning Electron Microscopy (SEM) picture of biologically produced Birnessite (top), and
SEM picture of physico-chemically produced Birnessite (bottom)
(photos: Arie Zwijnenburg, Wetsus)
54
4
BIOLOGICAL AND PHYSICO-CHEMICAL
FORMATION OF BIRNESSITE DURING RIPENING OF
MANGANESE REMOVAL FILTERS
Jantinus H. Bruins, Branislav Petrusevski, Yness M. Slokar, Koen Huysman, Koen Joris, Joop C.
Kruithof, Maria D. Kennedy(2015). Biological and physicochemical formation of Birnessite during the
ripening of manganese removal filters. Water Research, 69C: 154-161
55
4.1 Abstract
The efficiency of manganese removal in conventional groundwater treatment consisting of aeration
followed by rapid sand filtration, strongly depends on the ability of filter media to promote auto-catalytic
adsorption of dissolved manganese and its subsequent oxidation. Earlier studies have shown that the
compound responsible for the auto-catalytic activity in ripened filters is a manganese oxide called Birnessite.
The aim of this study was to determine if the ripening of manganese removal filters and the formation of
Birnessite on virgin sand is initiated biologically or physico-chemically. The ripening of virgin filter media
in a pilot filter column fed by pre-treated manganese containing groundwater was studied for approximately
600 days. Samples of filter media were taken at regular time intervals, and the manganese oxides formed in
the coating were analysed by Raman spectroscopy, Electron Paramagnetic Resonance (EPR) and Scanning
Electron Microscopy (SEM). From the EPR analyses, it was established that the formation of Birnessite
was most likely initiated via biological activity. With the progress of filter ripening and development of the
coating, Birnessite formation became predominantly physico-chemical, although biological manganese
oxidation continued to contribute to the overall manganese removal. The knowledge that manganese
removal in conventional groundwater treatment is initiated biologically could be of help in reducing typically
long ripening times by creating conditions that are favourable for the growth of manganese oxidizing
bacteria.
Keywords: Manganese removal; Ripening time; Manganese oxides; Birnessite; Biological manganese
oxidation; Physico-chemical manganese oxidation
4.2 Introduction
As mentioned before, in Europe, the removal of manganese from groundwater for water supply is
commonly achieved through conventional aeration-rapid sand filtration. This treatment is not only cost
effective but also environmentally friendly, because no chemicals (KMnO4, O3, Cl2) are required for
oxidation of Mn2+. The efficiency of manganese removal in aeration-filtration treatment of groundwater,
and particularly during ripening of virgin filter media, strongly depends on the ability of filter media to
adsorb dissolved manganese. Virgin quartz sand and anthracite, the most commonly used filter media, do
not have significant capacity to adsorb dissolved manganese. According to the redox potential - pH diagram
for aqueous manganese, chemical formation of MnOx requires a high redox potential and pH (Stumm and
Morgan, 1996). In general, in rapid sand filters used for groundwater treatment, conditions for chemical
formation of auto-catalytically active manganese oxides on virgin filter media, which can adsorb and
subsequently oxidize dissolved manganese, are poor since most groundwaters have a low redox potential
and pH.
MnOx-coated filter media could promote manganese removal through physico-chemical auto-catalytical
adsorption and subsequent oxidation of adsorbed manganese, as described by the oxidation kinetics for
dissolved Mn2+ by oxygen in aqueous solution (Stumm and Morgan, 1996):
56
-d[Mn2+]/dt
k0
=
=
-k0.[Mn2+] + k1.[Mn2+].[MnOX]
k.PO2.[OH-]2
Where:
k0
k1
=
=
reaction rate constant (L/min)
reaction rate constant (L/mol.min)
PO2
=
partial pressure of oxygen (atm)
(1)
(2)
Equation (1) shows that physico-chemical Mn2+ oxidation is enhanced by heterogeneous autocatalytic
activity of solid manganese oxides (e.g., MnOX) present in filter media coating. Furthermore equation (2)
shows that the reaction rate constant k0 is influenced by the oxygen concentration, and is also strongly pH
dependent. On the other hand, biological formation of manganese oxide is less pH dependent. Burger et.al.
(2008b) reported biological manganese removal at a pH of 6.5, whereas Hoyland et. al. (2014) found
biological manganese removal at a pH as low as 6.3. Several researchers investigated the influence of
different types of bacteria on manganese oxidation, amongst others, Leptothrix sp. (Adams and Ghiorse,
1985; Barger et al., 2009; Boogerd and De Vrind, 1987; Burger et al., 2008a, 2008b; Corstjens et al., 1997; El
Gheriany et al., 2009; Hope and Bott, 2004; Tebo et al., 2004, 2005), Pseudomonas sp. (Barger et al., 2009;
Brouwers et al., 1999; Caspi et al., 1998; DePalma, 1993; Gounot, 1994; Tebo et al., 2004, 2005;, Villalobos
et al., 2003, 2006) and Bacillus sp. (Barger et al., 2005, 2009; Brouwers et al., 2000; Mann et al., 1988; Tebo et
al., 2004, 2005). Furthermore several studies suggested that the manganese removal in rapid sand filters
could be accelerated by the use of ‘bio’-aged or coated filter media (Buamah et al., 2008, 2009a, 2009b; Hu
et al., 2004a, 2004b; Islam et al., 2010; Katsoyiannis and Zouboulis, 2004; Kim et al., 2009; Sahabi et al., 2009;
Tiwari et al., 2007). Therefore literature strongly suggests the impact of biology on manganese removal.
Bruins et al. (2014a) established that the manganese oxide responsible for the autocatalytic action in ripened
manganese removal filters is a mineral called Birnessite. Despite assumptions proposed in several literature
references that manganese removal is a biological process (Burger et al., 2008b; Vandenabeele et al., 1992;
Tekerlekopoulou et al., 2008), Bruins et al. (2014a) showed that Birnessite present in coatings of ripened
manganese removal filters (in operation for over 15 years) was of physico-chemical origin. However, their
results do not exclude that initially Birnessite layers were formed biologically.
The main focus of this research was to establish how the filter media ripening process, regarding manganese
removal in groundwater treatment based on aeration and rapid sand filtration was initiated. The objective
of the study was to show whether the MnOx (Birnessite) formed during the filter ripening was of biological
and/or physico-chemical origin.
The ripening process was studied in a pilot filter column (A1) located at the full-scale groundwater treatment
plant (GWTP) Grobbendonk, water supply company Pidpa, Belgium.
At GWTP Grobbendonk groundwater is treated containing iron, manganese and ammonia. The full-scale
groundwater treatment plant at this location consists of (cascade) aeration, filtration, post aeration (cascade)
with pH correction and post filtration. In the first filtration stage, iron is highly efficiently removed, while
partial ammonia removal is achieved. In the second filtration stage the remaining ammonia and most of the
manganese is removed. Filter media ripening of post filters with respect to manganese removal at this
57
location is a very fast process, and is typically completed within 14 to 21 days. To simulate post filtration of
the full scale GWTP, the pilot filter column was fed with water after post aeration with pH correction.
The filter column (A1) used in this study had an inner diameter of 10 cm and was filled with virgin quartz
sand of size fraction 0.7-1.25 mm, with a height of 0.30 m. The filter column was operated in down flow
mode at a filtration rate of 5.1 m/h with an empty bed contact time of 3.5 minutes. Backwashing was carried
out after approximately every 2 weeks of continuous operation. Backwashing was performed with water
only, at a backwash rate of 35 m/h, resulting in a filter bed expansion of approximately 20%.
To evaluate the ripening process of the filter media with respect to manganese removal and the formation
of MnOx, both water (feed, filtrate and backwash water), and filter media samples were collected at certain
time intervals. Filter media samples were taken from the top of the filter bed. Samples of backwash water
were allowed to settle, and settled particles were dried at room temperature.
To determine the pathway of Birnessite formation on the virgin sand during the ripening process and on
the coated media after prolonged filter runs, and to characterize, identify and classify the manganese oxide
in filter media coating, the following analytical methods were used:
-
Raman spectroscopy;
Electron paramagnetic resonance (EPR);
Scanning electron microscopy (SEM).
The techniques, instruments and sample preparation methods applied are described by Bruins et al. (2014a).
The obtained results were compared with references from literature and with a mineral Birnessite,
originating from the Aufgeklärt Glück Mine Hasserode near Wernigerode, Harz, Saxony-Anhalt, Germany
(Witzke, 2012).
The performance of the pilot filter was monitored for a filter run of 20 months. In this paper, the time
needed to achieve very effective (> 95 %) manganese removal starting with virgin sand is called the ‘filter
ripening time’.
58
4.4.1 Analytical data pilot
Quality of feed water for the pilot column is given in Table 4.1.
Table 4.1: Quality of feed water to the pilot filter column
Parameter
unit
range
Manganese
mg/L
0.100 – 0.150
Iron*)
mg/L
0.03 - 0.1
Ammonium
mg/L
0 - 0.20
pH
7.5 - 7.9**)
Oxygen
mg/L
8 - 9.5
Turbidity
FTU
0.5 - 1.2
Redox potential
mV
+200 - +290
Temperature
10.5 - 12.5
⁰C
Hydrogen carbonate
mg/L
115 - 135
* ) mainly in oxidized (Fe 3 + ) form
* * ) most of the time pH ranged from 7.5 to 7.6
From Table 4.1 it is clear that during this pilot study water quality conditions (pH and redox potential)
enabled Mn2+ adsorption and subsequent oxidation on existing manganese oxide layers (Bruins et al., 2014b).
However it was doubtful if these water quality conditions were suitable to achieve a fast and effective startup of the filter ripening process by adsorption and subsequent oxidation on virgin filter media (Graveland,
1971; Stumm and Morgan 1996).
Fig. 4.1 plots acidity and redox potential of feed water from this study (black dot), in the redox-pH diagram
of aqueous manganese (Stumm and Morgan, 1996). Redox-pH conditions for aerated groundwater from a
number of selected Dutch full-scale GWTPs are also included in the graph (grey dots).
59
Figure 4.1: Electron activity (pε) or redox potential (E h in V) – pH diagram for aqueous manganese
(adopted from Stumm and Morgan, 1996), compared to feed water conditions of the pilot (black dot)
and groundwater of other GWTPs (grey dots).
From Fig. 4.1 it can be seen that the pH and redox potential of feed water to the pilot (black dot) is in the
transition zone of conditions for dissolved manganese and conditions theoretically required to form
MnCO3. At a given pH the amount of dissolved Mn2+ in groundwater is controlled by the concentration of
hydrogen carbonate present in the water (Buamah, 2009b). Given the composition of the feed water to the
pilot filter column (Table 4.1), significant formation of manganese carbonate was not likely. This assumption
was confirmed by Raman spectroscopy analyses, by which no or negligible amounts of MnCO3 were found
in the coating. However, MnCO3 particles could be formed in limited amounts in the pilot feed water, after
pH correction (dosing of milk of lime, applied at the full-scale GWTP Grobbendonk upstream of the intake
point for the pilot, with temporarily locally a high pH). Consequently, the presence of small amounts of
MnCO3 particles deposited on the (virgin) filter media in the filter column of the pilot cannot be excluded.
For the ripening process this might be a disadvantage, because MnCO3 has no auto-catalytical adsorption
and oxidation properties, unlike manganese oxides (MnOX). Furthermore, MnCO3 precipitate can cover
already formed (auto-catalytically active) MnOX (Graveland, 1971; Graveland and Heertjes, 1975). Hence,
in this way MnCO3 is able to hinder the ripening process by retarding and disturbing the auto-catalytical
adsorption and oxidation reaction. Consequently, possible formation of MnCO3 and its retention on filter
media will not be beneficial for the process of filter media ripening. During this study, however, the negative
effect of MnCO3 was limited, because in line with the MnCO3 solubility (Buamah, 2009b) no or negligible
amounts of MnCO3 were identified with Raman spectroscopy in the coating.
4.4.2 Ripening time of filter media
Fig. 4.2 gives an overview of the manganese removal efficiency during the first 25 days of the ripening
process of the pilot column. From Fig. 4.2 it can be seen that ripening of the filter media (sand) was
completed after approximately 25 days.
60
A ripening period of approximately 25 days is relatively short. Although in some situations low Mn2+
concentration may lead to a relatively short ripening time, in general, it typically takes more than two to
three months, and in some exceptional cases even more than a year, to achieve complete manganese removal
even with low Mn2+ concentrations (Bahlman, 2014; Cools, 2010; Krull, 2010).
removal efficiency Mn (%)
100
80
60
40
20
0
0
5
10
15
20
25
days after start up
Column A1 (sand)
Figure 4.2: Ripening period of pilot column A1 (height 0.3 m, diameter 0.1 m), filtration rate
5.1 m/h.
The build-up of MnOX layers on virgin filter media can be seen using a microscope. Fig. 4.3 shows images
of virgin sand and the same filter media after a filter run time of 81 days. The observed thin grey layer with
(tiny) blackish spots (Fig. 4.3 (R)) represents MnOx deposits.
Figure 4.3: Virgin sand (L) and sand from column A1 after a filter run time of 81 days (R).
Microscope, magnification 20x.
4.4.3 Raman spectroscopy
Raman spectra of a mineral Birnessite reference, MnOX deposits on filter media (from pilot column A1)
18 days after initiating filter media ripening, and MnOX in solids from the backwash water (BW) from the
pilot column (28 days after initiation of filter media ripening) were compared. From Fig. 4.2 it can be seen
that the sand sample from the pilot column was taken at the stage of exponential increase of manganese
removal efficiency. A few hours before sampling the coated sand, the manganese removal efficiency was
61
approximately 25%, while two days later the manganese removal efficiency was increased to approximately
65% (Fig. 4.2). Consequently, the sand sample was taken at a very early stage in the ripening process at
which the first manganese oxides were formed.
From Raman spectra comparison it was found that the main peaks obtained for solids from backwash water
and sand from the pilot column were comparable with the Raman shift, of the main peaks of the Birnessite
reference (Witzke, 2012). The positions of the peaks found in this study were also in accordance with the
Birnessite peaks found in literature (Ma et al., 2007). Consequently, this strongly suggests that manganese
oxide formed on the filter media already at the early stage in the ripening process was Birnessite.
After a number of filter run times Raman spectra were analysed for several filter media samples and solids
from backwash water (Fig. 4.4). For these samples the main peaks were always found at comparable Raman
shifts. From this it can be concluded that during the complete filter run time, the manganese oxide found
in the filter media coating and the solids in the backwash water was Birnessite.
A
200
B
300
400
500
600
700
800
900
200
300
Raman shift (cm-1)
A1 (18)
A1 (128)
A1 (224)
400
500
600
700
800
900
Raman shift (cm-1)
A1 (425)
BW(18)
BW(128)
BW(273)
BW(425)
Figure 4.4: Raman spectra of manganese oxide for different filter media samples from the pilot column
(A1) and different samples of solids from backwash water (BW), both taken at different filter run
times.
Observed differences in the height of peaks for different samples (Fig. 4.4) can likely be attributed to
different concentrations of counter ions (e.g., sodium, calcium, etc.), present in the lattice of the manganese
oxide, as reported by Julien et al. (2003, 2004). The presence of other ions was confirmed by ICP-MS and
SEM-EDX analysis. Although the chemical formula(s) of Mn-oxides belonging to the Birnessite group can
be expressed in different ways, they always include several counter ions (e.g., (Na+,Ca2+,Mn2+)Mn7O14
2.8H2O, Post, 1999).
From the Raman spectroscopy analyses it can be concluded that manganese oxide in the filter media coating,
formed during the ripening, and the solids present in the filter backwash water, were a Birnessite type of
manganese oxide. In addition it can be concluded that Birnessite was already present at a very early stage of
the filter ripening process. Finally, the peaks in the Raman spectra of both the filter media coating and solids
from the backwash water were not very sharp, indicating that the Birnessite is poorly crystalline or
amorphous.
4.4.4 Electron paramagnetic resonance (EPR)
Kim et al. (2011) recently reported that electron paramagnetic resonance (EPR) spectra could distinguish
biologically formed Birnessite from inorganic (abiogenic) Birnessite. Especially the measured wavelength,
expressed in gauss, was reported to be a criterion to distinguish between the two sources of Birnessite. Kim
62
et al. (2011) made the following classification based on wavelength (ΔH) in gauss (measured at room
temperature - 290 K):
Birnessite of bacterial origin
Birnessite of biomineral origin *)
Birnessite of abiogenic origin
: ΔH
: ΔH
: ΔH
< 600
600-1200
> 1200
gauss
gauss
gauss
*)biomineral = a mineral (Mn nodules from fresh and marine water and desert varnish) of possible
biological origin, that has been transformed over a long time into an abiogenic mineral.
Fig. 4.5 shows the EPR spectrum of the Birnessite reference (Witzke, 2012) used in this study. The
calculated wavelength (ΔH) obtained from this Birnessite reference is significantly larger than 1200 G,
confirming its abiogenic (inorganic) origin.
ΔH > 1200
0
1000
2000
3000
4000
5000
6000
7000
gauss
Birnessite (reference)
Figure 4.5: EPR spectrum at 290 K of the Birnessite reference.
EPR spectra were also measured for the collected samples of (coated) sand from the filter column and
samples of solids from the backwash water. In Fig. 4.6 the calculated wavelengths (ΔH in gauss) of the EPR
spectra of coated sand samples (red diamonds), and the backwash water samples (green triangles) were
plotted as a function of the filter run time (in days). The blue dot in Fig. 4.6 shows the ΔH value of Birnessite
present in the filter media coating of a full-scale filter of the 2nd filtration stage at GWTP Grobbendonk,
after a filter run of at least 4 years. This full-scale filter received the same feed water as the pilot column.
63
ΔH (gauss)
1600
1200
800
400
0
0
100
200
300
400
500
600
Filter run time (days)
Column A1
Full scale plant
Backwash water
Figure 4.6: Wavelengths (ΔH in gauss) of EPR Birnessite spectra from coated sand samples from the
pilot column (red diamonds), and solid samples from the backwash water (green triangles), versus filter
run time.
From Fig. 4.6 it can be seen that the ΔH values of the Birnessite present in the coating of filter media
increased with the filter run time. ΔH values at the start of the filter operation were just above the fine
dotted line representing the boundary between Birnessite of biological and biomineral origin. In this study,
values in the biomineral range might indicate a mixture of biological and abiogenic Birnessite. With a
prolonged filter run time, ΔH values increased up to values above the coarse dotted line, reflecting the
presence of Birnessite of physico-chemical origin. The increase in ΔH can be explained by the physicochemical manganese adsorption/oxidation reaction due to the presence of formed manganese oxides on
the media surface. It is known that the physico-chemical manganese adsorption/oxidation reaction is an
extremely fast auto-catalytic process (Graveland, 1971; Stumm and Morgan, 1996). Hence, the amount of
physico-chemically formed Birnessite on the filter media may increase compared to the amount of
biologically formed Birnessite.
A rapid increase in ΔH values was observed initially. Given the exponential shape of the ΔH curve, and the
fact that the first sample was taken 18 days after the start of the filter run, it is very likely that the initial
values of ΔH were even below the value of 600 G proposed by Kim et al. (2011) for Birnessite of biological
origin. Samples of material from the coating taken after a prolonged filter run time showed ΔH values
approaching that of Birnessite of abiogenic origin. Samples of a coating from an aged full-scale filter showed
a ΔH value representing Birnessite of abiogenic origin (ΔH > 1200 G).
The average ΔH value of Birnessite present in solid samples collected from the backwash water was 609 G,
with a standard deviation of 111 G. This suggests that Birnessite found in filter backwash water was
predominantly of biological origin. ΔH values of Birnessite found on solids in backwash water were
consistently lower than values found for Birnessite present in the filter media coating throughout the whole
testing period of approximately 600 days.
EPR results suggest that physico-chemical formation of manganese oxide on filter media is the predominant
process, once initial layers of biological Birnessite are formed. At the same time, results of the EPR analysis
of solids found in the filter backwash water indicate that biological oxidation of Mn2+ takes place throughout
the whole filter run time. Biologically formed Birnessite in solids found in backwash water are likely formed
64
on and around the bacterial cells attached on the surface of the filter media. This biomass present on filter
grains with formed Birnessite solids is partially flushed out during backwashing. Backwashing of the pilot
column is done with water only, with filter bed expansion ≤20%, which is not very abrasive, and hence, it
can be assumed that only a limited amount of Birnessite of physico-chemical origin, that is strongly attached
to filter media, will be flushed out.
Results emerging from this EPR study strongly suggest that filter media ripening with respect to manganese
removal starts biologically, and at a later stage is controlled by the physico-chemical oxidation of adsorbed
Mn2+.
The conclusions of this study based on the manganese oxide identification and characterization by Raman
spectroscopy, EPR and SEM can be further underpinned by showing the presence of manganese oxidizing
bacteria. Therefore molecular techniques such as qPCR, Malditof and next generation DNA sequencing
were applied. With these techniques the presence of bacteria generally known to be able to oxidize Mn2+
such as Pseudomonas sp. and Leptothrix sp. were confirmed. These results confirm the conclusions based on
Raman spectroscopy, EPR and SEM measurements of manganese oxides in filter coatings described in this
paper. The detailed identification of manganese oxidizing bacteria and their relevance for manganese
removal in practice is a comprehensive study, which will be published in a separate paper (Bruins et al., to
be published in 2015).
4.4.5 Scanning electron microscopy (SEM)
To examine the structural difference between biologically and physico-chemically formed Birnessite,
samples of filter media taken after different filter run times were analysed by SEM. Jiang et al. (2010) showed
a distinctive difference in structure between biologically (Pseudomonas putida) and physico-chemically formed
Birnessite by SEM imaging. Biologically produced Birnessite has a fluffy plate structure in contrast to
physico-chemically formed Birnessite that has a more coral or sponge type structure. The SEM images of
filter media coating from this study are compared with the findings of Jiang et al. (2010).
Fig. 4.7 shows SEM micrographs of Birnessite found in the coating of the filter media after different run
times.
Figure4.7: SEM image of Birnessite found in the coating of the filter media (sand) after 128 days
(left), 394 days (middle) and 273 days (right) of filter run time (magnification resp. 5,000x; 5,000x;
10,000x).
Fig. 4.7 (left) shows the presence of fluffy plate structures, typical for Birnessite of biological origin (Jiang
et al., 2010). The SEM micrograph of aged filter media (Fig. 4.7, middle), shows coral or sponge-like
structures, typical for Birnessite of physico-chemical origin (Jiang et al., 2010).
65
Fig. 4.7 (right) shows a SEM micrograph of Birnessite present in aged filter media made at a larger
magnification (10,000x). Structures typical for physico-chemically formed Birnessite are clearly visible e.g.,
‘layered’ patterns within the coral or sponge structure.
According to Post (1999), Birnessite consists of stacked octahedral MnO6 layers, with different interlayer
cations and water. According to Post (1999) such a structure, allows Birnessite to ‘readily participate in
oxidation-reduction and cation-exchange reactions, and therefore plays a significant role in soil and
groundwater chemistry’. This consequently makes Birnessite extremely suitable for the auto-catalytic
oxidation reaction in conventional aeration-filtration GWTPs.
The SEM images also confirmed that the Birnessite is poorly crystalline or amorphous.
The character of manganese removal being biological or physico-chemical has been debated for quite some
time. Some researchers propose a physico-chemically driven process (Graveland and Heertjes, 1975),
whereas others suggest a biological process (Tekerlekopoulou et al., 2008; Burger et al., 2008a, Vandenabeele
et al., 1992). When strong oxidizers, such as chlorine, ozone and potassium permanganate are used, it is
obvious that the formed MnO2 (Pyrolusite) is produced in a physico-chemical way (Knocke et al., 1991).
However, when groundwater treatment by conventional aeration-filtration, without the use of strong
oxidants, is applied the common quality of groundwater (relatively low pH, low redox potential) is, in
general, not supportive for a fast start of virgin filter media ripening through a physico-chemical mechanism
(Bruins et al., 2014b). Burger et.al.(2008b) and Hoyland et. al. (2014) have shown that contrary to chemical
oxidation, biological oxidation of Mn2+ can take place at low pH. The current study shows that biological
Mn2+ oxidation plays a decisive role in filter media ripening. After prolonged time the formed manganese
oxide Birnessite becomes of a more physico-chemical origin, although part of the Birnessite still has a
biological origin. However, results emerging from this study were insufficient to clarify if the biological
oxidation of manganese remains the controlling manganese removal mechanism also after initial ripening
of the filter media. Although substantial research has been performed on this topic, little is known why
heterotrophic bacteria oxidize manganese and suggested reasons are at least speculative (De Schamphelaire
et al., 2007). Therefore, it is important to unravel how heterotrophic bacteria (e.g., Pseudomonas sp., Leptothrix
sp.), also found in this research, can benefit from this oxidation reaction. Furthermore the question must
be answered if Pseudomonas sp. and Leptothrix sp. are acting alone or in cooperation with other bacteria.
Understanding the biochemical and physico-chemical mechanisms by which manganese coating of filter
media is formed during the filter ripening could help creating conditions favouring Birnessite formation.
This knowledge together with the results of this study may help reducing long ripening periods of manganese
removal filters with virgin filter media, making groundwater treatment by conventional aeration-filtration
more generally applicable.
4.5 Conclusions
The following conclusions can be drawn based on the results emerging from this study:
•
•
Analysis of the filter media from the manganese removal plant conducted with Raman
spectroscopy, Electron Paramagnetic Resonance (EPR) and Scanning Electron Microscopy (SEM)
showed that a Birnessite type of manganese oxide is the predominant mineral in the coating;
Raman spectroscopy results showed that Birnessite is already present in the coating at a very early
stage of the ripening process;
66
•
•
•
•
From the EPR analyses and comparison with literature, the Birnessite type of manganese oxide, at
the beginning of the ripening process was most likely of biological origin. Over a prolonged filter
run time, Birnessite changed from a predominantly biologically formed to a physico-chemically
formed compound;
Solids collected from filter backwash water throughout the whole ripening period were consistently
of biological origin, suggesting that biological oxidation of adsorbed manganese was present
throughout the filter run, and contributed to manganese removal;
SEM micrographs showed a clear difference between biologically and physico-chemically formed
Birnessite. Biologically produced Birnessite is fluffy, plate structured, whereas physico-chemically
produced Birnessite shows more a sponge or coral structure;
Understanding the mechanisms by which a manganese coating of filter media starts up could
endorse the creation of conditions favouring Birnessite formation, and possibly help in reducing
typically long ripening periods of manganese removal filters with virgin filter media.
This research was financially and technically supported by WLN and the Dutch water supply companies
Groningen (WBG) and Drenthe (WMD). The authors are also grateful to Dr. Arie Zwijnenburg and Mr.
Ton van der Zande (Wetsus) for providing the SEM and Raman spectroscopy analyses and to Dr. Wesley
Browne (University of Groningen) for giving access to the Electron Paramagnetic Resonance
measurements. Also thanks to Dr. Witzke for providing a mineral Birnessite (reference) sample. Last but
certainly not least, thanks go to the water company Pidpa (Belgium) for their willingness to share the data
from their groundwater treatment plant with us and to give us the opportunity to perform a pilot test at
Grobbendonk. We especially thank Mrs. Ann Maeyninckx and Mrs. Martine Cuypers (Pidpa) for fulfilling a
great task by running the pilot and performing in situ analyses.
4.7 References
Adams, L.F., Ghiorse, W.C., 1985. Influence of Manganese on Growth of a Sheathless Strain of Leptothrix
discophora. Applied and Environmental Microbiology, 49 (3), 556 –562.
Bahlman, J.A., 2014. Evides Waterbedrijf (Dutch water company), personal communication, The
Netherlands.
Barger, J.R., Tebo, B.M., Bergmann, U., Webb, S.M., Glatzel, P., Chiu, V.Q., Villalobos, M., 2005. Biotic
and abiotic products of Mn (II) oxidation by spores of the marine Bacillus sp. strain SG-1. American
Barger, J.R., Fuller, C.C., Marcu, M.A., Brearly, A., Perez De la Rosa M., Webb, S.M., Caldwell, W.A., 2009.
Boogerd, F.C., De Vrind, J.P.M., 1987. Manganese oxidation by Leptothrix discophora. Journal of Bacteriology,
489-494.
67
Brouwers, G.J., De Vrind, J.P.M, Corstjens, P.L.A.M., Cornelis, P., Baysse, C., De Vrind-De Jong, E.W.,
1999. CumA, a Gene Encoding a multicopper oxidase, is involved in Mn2+ oxidation in Pseudomonas
putida GB-1. Applied and Environmental Microbiology, 65(4), 1762-1768.
Brouwers, G.J., Vijgenboom, E., Corstjens, P.L.A.M., De Vrind, J.P.M., De Vrind-De Jong, E.W., 2000.
Bacterial Mn2+ oxidizing systems and multicopper oxidases: an overview of Mechanisms and
Functions. Geomicrobiology Journal, 17(1), 1-24.
Bruins, J.H., Petrusevski, B., Slokar, Y.M., Kruithof, J.C., Kennedy, M.D., 2014a. Manganese removal from
groundwater: characterization of filter media coating. Desalination and Water Treatment (in press,
on line 18 June 2014, DOI:10.1080/19443994.2014.927802).
Bruins, J.H., Vries, D., Petrusevski, B., Slokar, Y.M., Kennedy, M.D., 2014b. Assessment of manganese
removal from over 100 groundwater treatment plants. Journal of Water Supply: Research and
Technology-AQUA, 63.4 268-280.
Buamah, R., Petrusevski, B., Schippers, J.C., 2008. Adsorptive removal of manganese (II) from the aqueous
phase using iron oxide coated sand. Journal of Water Supply: Research and Technology-AQUA 57.1,
1-11.
Buamah, R., Petrusevski, B., de Ridder, D., Van de Watering and Schippers, J.C., 2009a. Manganese removal
Technology: Water Supply 9.1: 89 - 98.
Buamah, R., 2009b. Adsorptive removal of manganese, arsenic and iron from groundwater. PhD thesis,
Unesco-IHE, Delft, Wageningen University, The Netherlands.
Burger, M.S., Krentz, C.A., Mercer, S.S., Gagnon, G.A., 2008a. Manganese removal and occurrence of
Technology-AQUA 57.5.
Burger, M.S., Mercer, S.S., Shupe, G.D., Gagnon, G.A., 2008b. Manganese removal during bench-scale
biofiltration, Water Research 42, 4733-4742.
Caspi, R., Tebo, B.M., Haygood, M.G., 1998. C-Type Cytochromes and Manganese Oxidation in Pseudomonas
putida MnB1. Applied and Environmental Microbiology, 64 (10), 3549-3555.
Cools, B., 2010. De Watergroep (Flemish water company), personal communication, Belgium.
Corstjens, P.L.A.M., De Vrind, J.P.M., Goosen, T., De Vrind-de Jong, E.W., 1997. Identification and
molecular analysis of the Leptothrix discophora SS-1 mofA gene, a gene putatively encoding a
manganese-oxidizing protein with copper domains. Geomicrobiology Journal, 14 (2), 91-108.
DePalma, S.R., 1993. Manganese oxidation by Pseudomonas putida. PhD thesis, Harvard University,
Cambridge, USA.
68
De Schamphelaire, L. , Rabaey, K. , Boon, N. , Verstraete, W. And Boeckx, P. 2007. Minireview: The
potential of enhanced manganese redox cycling for sediment oxidation. Geomicrobiology Journal,
24 (7), 547-558.
El Gheriany, I.A., Bocioaga, D., Hay A.G, Ghiorse, W.C., Shuler, M.L., Lion, L.W., 2009. Iron Requirement
for Mn (II) Oxidation by Leptothrix discophora SS-1. Applied and Environmental Microbiology, 75 (5),
1229-1235.
Gounot, A-M., 1994. Microbial oxidation and reduction of manganese: Consequences in groundwater and
applications. FEMS Microbiology Reviews 14, 339-350.
Graveland, A., 1971. Removal of manganese from groundwater. PhD thesis, Technical University Delft,
The Netherlands.
Graveland, A., Heertjes, P.M., 1975. Removal of manganese from groundwater by heterogeneous
autocatalytic oxidation. Trans. Instn. Chem. Engrs., 53, 154-164.
Hope, C.K., Bott, T.R., 2004. Laboratory modelling of manganese biofiltration using biofilms of Leptothrix
doscophora, Water Research 38, 1853-1861.
Hoyland, V.W., Knocke, W.R., Falkinham III, J.O., Prude, A., Singh, G., 2014. Effect of drinking water
process treatment parameters on biological removal of manganese from surface water, Water
Research, in press (on line since Augustus 2014).
Hu, P-Y., Hsieh, Y-H., Chen, J-C., Chang, C-Y., 2004a. Adsorption of divalent manganese ion on
manganese-coated sand. Journal of Water Supply: Research and Technology-AQUA 53.3, 151-158.
Hu, P-Y., Hsieh, Y-H., Chen, J-C., Chang, C-Y., 2004b. Characteristics of manganese-coated sand using
SEM and EDAX analysis. Journal of Colloid and Interface Science, 272, 308-313.
Islam, A.A., Goodwill, J.E., Bouchard, R., Tobiasen, J.E., Knocke, W.R., 2010. Characterization of filter
media MnO2(s) surfaces and Mn removal capability. Journal AWWA, 102 (9), 71-83.
Jiang, S., Kim, D-G., Kim, J., Ko, S-O., 2010. Characterization of the biogenic manganese oxides produced
by Pseudomona putida strain MnB1. Environ. Eng. Res., 15 (4), 183-190.
Julien, C., Massot, M., Baddour-Hadjean, R., Franger, S., Bach, S., Pereira-Ramos, J.P., 2003. Raman spectra
of birnessite manganese dioxides. Solid State Ionics 159, 345– 356.
Julien, C., Massot, M., Poinsignon, C., 2004. Lattice vibrations of manganese oxides Part I. Periodic
structures. Spectrochimica Acta Part A 60, 689-700.
Katsoyiannis, I. A., Zouboulis A.I., 2004. Biological treatment of Mn (II) and Fe (II) containing
groundwater: kinetic considerations and product characterization. Water Research 38, 1922-1932.
Kim, W.G., Kim, S.J., Lee, S.M., Tiwari, D., 2009. Removal characteristics of manganese-coated solid
samples for Mn (II). Desalination and Water Treatment 4, 218-223.
69
Kim, S.S., Bargar, J.R., Nealson, K.H., Flood, B.E., Kirschvink, J.L., Raub, T.D., Tebo, B.M., Villalobos,
M., 2011. Searching for Biosignatures Using Electron Paramagnetic Resonance (EPR) Analysis of
Manganese Oxides. Astrobiology, 11(8), 775-786.
Knocke, W.R., Van Benschoten, J.E., Kearny, M.J., Soborski, A.W. & Reckhow, D.A. 1991. Kinetics of
Manganese and Iron oxidation by Potassium Permanganate and Chlorine dioxide. Journal of AWWA,
June, 80-87.
Krull, J., 2010. Stadtwerke Emden (German water company), personal communication, Germany.
Ma, S-B., Ahn, K-Y., Lee, E-S., Oh, K-H., Kim, K-B., 2007. Synthesis and characterization of
manganese dioxide spontaneously coated on carbon nanotubes. Carbon 45, 375-382.
Mann, S., Sparks. N.H.C., Scoot, G.H.E., De Vrind-De Jong E.W., 1988. Oxidation of Manganese and
formation of Mn3O4 (Hausmannite) by spore coats of a marine Bacillus sp. Environmental
Microbiology, 54 (8), 2140-2143.
Post, J. E., 1999. Manganese oxide minerals: Crystal structures and economic and environmental
significance. Proceedings of the National Academy of Sciences USA, 96, 3447-3454.
Sahabi, D.M., Takeda, M., Suzuki, I., Koizumi, J-I., 2009. Removal of Mn2+ from water by “aged” biofilter
media: The role of catalytic oxides layers. Journal of Bioscience and Bioengineering, 107(2), 151-157.
Stumm, W. and Morgan, J.J., 1996. Aquatic Chemistry, chemical equilibria and rates, 3rd ed. Wiley, New
York.
Tebo, B.M., Marger, J.R., Clement, B.G., Dick, G.J., Murray, K.J., Parker, D., Verity, R., Webb, S.M., 2004.
32, 287-328.
Tebo, B.M., Johnson, H.A., McCarthy, J.K., Templeton, A.S., 2005. Geomicrobiology of manganese (II)
oxidation. TRENDS in Microbiology, 13(9), 421-428.
Tekerlekopoulou, A.G., Vasiliadou, I.A., Vayenas, D.V., 2008. Biological manganese removal from potable
water using trickling filters. Biochemical Engineering Journal 38, 292-301.
Tiwari, D., Yu, M.R., Kim, M.N., Lee, S.M., Kwon, O.H., Choi, K.M., Lim, G.J., Yang, J.K., 2007. Potential
Vandenabeele, J., De Beer, D., Germonpré, R., Verstreate, W., 1992. Manganese oxidation by microbial
consortia from sand filters. Microbial Ecology 24, 91-108.
Villalobos, M., Toner, B., Bargar, J., Sposito, G., 2003. Characterization of the manganese oxide produced
by Pseudomonas putida strain MnB1. Geochimica et Cosmochimica Acta, 67(14), 2649-2662.
Villalobos, M., Lanson, B., Manceau, A., Toner, B., Sposito, G., 2006. Structural model for the biogenic Mn
Witzke, T., 2012. Mineralogist PANanalytical, personal communication, Germany.
70
71
Figure: Pseudomonas grimontii (from top to bottom; magnification 5,000x | 10,000x | 20,000x Photos made by Jelmer Dijkstra (Wetsus), sample preparation by Pim Willemse (WLN).
72
5
IDENTIFACTION OF THE BACTERIAL POPULATION IN
MANGANESE REMOVAL FILTERS
Main part of this chapter was submitted as:
Jantinus H. Bruins, Branislav Petrusevski, Yness M. Slokar, Gerhard H. Wübbels, Koen Huysman,
Bart A. Wullings, Koen Joris, Joop C. Kruithof, Maria D. Kennedy(2016). Identification of the
bacterial population in manganese removal filters. Submitted to Water Science and Technology: Water
Supply.
73
5.1 Abstract
Rapid ripening of manganese removal filters with virgin filter media, in conventional aeration-rapid sand
filtration treatment of groundwater, greatly depends on the rate of formation of adsorptive MnOx on filter
media. Earlier studies have shown that the manganese oxide responsible for autocatalytic manganese
removal in ripened filters is biologically formed Birnessite. The aim of this study was to identify bacteria
present in freshly ripened filters for manganese removal. Samples of backwash water were taken from 1st
stage (iron removal) filters and the freshly ripened manganese removal filter. The bacterial population was
identified with ‘next generation’ DNA sequencing, and specific bacteria were quantified with qPCR and
characterized by MALDI-TOF MS analysis.
The ‘next generation’ DNA sequencing analysis showed a bacteria population shift from the iron oxidizing
species Gallionella sp. in the iron removal filter to manganese and nitrite oxidizing species Pseudomonas sp.
and Nitrospira sp., respectively present in the manganese removal filter.
qPCR analysis confirmed the presence of a low concentration of the well-known Mn2+-oxidizing species P.
putida in the manganese removal filter backwash water.
Bacteria of the genus Pseudomonas, isolated from backwash water from the manganese removal filter were
cultured and identified with MALDI-TOF MS analysis. Amongst others, P. gessardii, P. grimontii, and
P. koreensis were identified.
The presence of high(er) concentrations of Pseudomonas spp. and minor amounts of P. putida and Leptothrix
sp. in the ripened filter media supports the assumption that a microbial consortium is involved in the
oxidation of manganese. However, inoculation of an isolated Pseudomonas species in a fermenter did not
result in the production of MnOx under the performed laboratory conditions, whereas P. putida strain
(ATCC 23483) was able to do so.
Keywords: biological manganese oxidation, Birnessite, manganese removal ripening time, molecular
DNA techniques, Pseudomonas sp.
5.2 Introduction
An important drawback of aeration-rapid sand filtration, commonly applied in several West European
countries to remove manganese from groundwater, is the long filter media ripening period. Farnsworth et
al. (2012) reported that manganese oxide formed on the filter media, responsible for manganese removal, is
a Birnessite type of mineral. Due to its structure, Birnessite has outstanding properties to adsorb and
subsequently oxidize Mn2+ (Post and Veblen, 1990; Post 1999). Bruins et al. (2015a) showed that Birnessite
present in the coating of a ripened manganese removing filter in operation for over 15 years, was of
physicochemical origin. Chemical formation of Birnessite requires alkaline conditions (Feng et al., 2005).
The redox potential - pH diagram for aqueous manganese (Stumm and Morgan, 1996), suggest that besides
a high pH, a high redox potential is required for chemical formation of MnOx. Such water quality
characteristics are not common for groundwater with, usually, low pH and low redox potential. Using
electron paramagnetic resonance and Raman spectroscopy, Bruins et al. (2015b), showed that formation of
Birnessite most likely starts through biological activity. In a number of other studies, it was also proposed
74
that manganese removal is an obligatory biological process (Burger et. al., 2008a,b, Farnsworth et al., 2012;
Katsoyiannis and Zouboulis, 2004; Vandenabeele et al., 1992; Tekerlekopoulou et al., 2008). Several species
of bacteria (Pseudomonas sp., Leptothrix sp. and Bacillus spores) able to-, or involved in oxidation of Mn2+ have
been identified in (ground)water (Tebo et al., 2005; Kim et. al., 2011). Pseudomonas sp., in particular Pseudomonas
putida, were extensively studied in relation to manganese oxidation (Barger et al., 2009; Brouwers et al., 1999;
Caspi et al., 1998; DePalma, 1993; Gounot, 1994; Tebo et al., 2004, 2005; Villalobos et al., 2003, 2006). Similar
studies were performed for Leptothrix sp. (Adams and Ghiorse, 1985; Barger et al., 2009; Boogerd and De
Vrind, 1987; Burger et al., 2008a, 2008b; Corstjens et al., 1997; El Gheriany et al. 2009; Hope and Bott, 2004;
Tebo et al. 2004, 2005) and Bacillus sp. spores (Barger et al., 2005, 2009; Brouwers et al., 2000; Mann et al.,
1988; Tebo et al., 2004, 2005; Vrind de, et al., 1986). Experiments with (spores of) Bacillus sp. were, however,
often done with species of marine origin (Mann et al., 1988; Webb et al., 2005a, 2005b).
Pseudomonas sp. and Leptothrix sp. are heterotrophic bacteria. Leptothrix sp. is able to oxidize iron as well as
manganese, whereas Pseudomonas sp. is able to oxidize manganese only (Daum et al., 1998; Fleming et.al,
2011). Despite extensive research, it is still not clear how heterotrophic bacteria benefit from manganese
oxidation (De Schamphelaire et al., 2007; Tebo et al. 2005; Geszvain et al., 2013). However, from literature
it is known that Mn2+ oxidation to MnO2, by Pseudomonas putida, could be beneficial for the co-metabolic
degradation of micro pollutants, resulting in the formation of easily accessible organic compounds for their
metabolism (Sabirova et al., 2008; Forrez et al., 2010; Meerburg et al., 2012). Similarly, it is proposed that
complex organic molecules (e.g., Natural Organic Matter (NOM), such as humic acids in groundwater)
undergo degradation by the same process, performed by e.g., Pseudomonas putida (Verstraete, 2013). Based on
the proposed model of Mn2+ oxidation by e.g., Pseudomonas putida and co-metabolic degradation of organic
micro pollutants from Meerburg et al. (2012), a simplified degradation scheme for NOM is adopted (Fig.
5.1). The process of co-metabolic degradation it is titled “bio cracking”.
Figure 5.1: Simplified co-metabolic degradation scheme of NOM by biological MnO x oxidation
(adopted from Meerburg et al.,(2012) and Verstraete (2013)).
From Fig. 5.1 it can be seen that MnOx, formed during the process of bio cracking, is reduced again to
Mn2+, in a self-supporting metabolic cycle. Literature suggested that not one bacterium is responsible for
manganese oxidation, but a microbial consortia (Vandenabeele et al., 1992; Vandenabeele, 1993; Verstraete,
2013). Once biological Birnessite (MnOx) is formed on filter media, it has extremely high adsorptive capacity
for metal ions, such as Mn2+ (Webb et al, 2005; Jiang et al., 2010). In this way biologically produced Birnessite
promotes and accelerates manganese removal through physicochemical autocatalytic adsorption and
75
subsequent oxidation of adsorbed manganese according to the oxidation kinetics of dissolved Mn2+ by
oxygen in aqueous solution (Stumm and Morgan, 1996).
The goal of this study was to provide additional insight in the role of biology in manganese removal. Specific
objective of the study was to identify (with molecular (DNA) techniques) species of bacteria present in iron
removal, and freshly ripened manganese removal filters. Furthermore the capability of selected bacterial
species found in manganese removal filters to oxidize Mn2+ was investigated in the laboratory. This study
will enhance knowledge of the role of biological activity in the ripening of manganese filters in practice and
show how to create conditions favorable for the biological manganese oxidation process.
The experiments presented in this study were carried out on a pilot plant located at full scale groundwater
treatment plant (GWTP) ‘Grobbendonk’, water supply company Pidpa, Belgium.
The full scale plant consists of a pre-aeration step (cascade), 1st filtration stage, post aeration (cascade) with
pH correction, and post filtration. GWTP Grobbendonk treats groundwater containing iron, manganese
and ammonia. Very effective removal of iron (> 98 %) is achieved in the 1st filtration step. Iron is removed
predominantly through adsorptive and biological mechanisms, with support of the bacterium Gallionella sp.
Filter media ripening of the post filters in this plant, concerning manganese removal, is a very fast process
(typically complete manganese removal is achieved within approx. 16 days). It is believed that the filter
media ripening process at this location is initial biologically. Similar to post filters in the full scale plant, the
pilot filter column was fed with re-aerated filtrate from the first filter stage of the full scale GWTP, after pH
correction (pH = 7.6). The pilot filter column is named in the rest of the paper as ‘column A1’. Details of
GWTP Grobbendonk and the pilot filter column are given in Bruins et al. (2015b).
To determine the presence and composition of the bacteria population, during the filter media ripening
process samples were taken from backwash water from the 1st filtration step in the full scale plant
Grobbendonk (‘BW 1st RSF’), and backwash water from the pilot filter column A1 ( ‘BW A1’). Backwash
water was used to obtain a higher bacteria yield. ‘BW A1’ were sampled when filter media ripening was
almost completed (i.e., manganese removal efficiency in the filter was > 90 %).
Measurements and subsequent identification of bacteria, present in ‘BW 1st RSF’ and ’BW A1’ were carried
out with “Next generation DNA sequencing”, quantitative Polymerase Chain Reaction (qPCR) and Matrixassisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS). Finally, growth
tests were performed in a fermenter with selected bacteria from a sample of ‘BW A1’, to examine if MnOx
could be produced biologically. In addition a pure culture of Pseudomonas putida (ATCC 23483, LMG 2321),
grown on a selective growth medium (section 5.3.4), was used as a reference.
5.3.1 Next generation DNA sequencing
Samples were taken from both, ‘BW 1st RSF’ and ‘BW A1’. A part of the 16S rRNA gene (approximately
900 bp) was amplified from these samples, using a eubacterial forward primer GM3 (5’-AGA GTT TGA
TCM TGG C-3’), and the universal reverse primer 926r (5’-CCG TCA ATT CMT TTG AGT TT-3’) with
identifiable sample bar codes. The pyrosequencing analysis of the amplified 16S rRNA genes was performed
using at LGC genomics (Berlin, Germany) with a 454 Life Sciences GS FLX series genome sequencer
upgraded to long read length (Roche, The Netherlands). The returned 16S rRNA gene sequences were
76
analysed, trimmed, aligned, and identified using the metagenomics tool in the software package Bionumercs
(Schloss et al., 2011) based on the 454 SOP (Schloss et al., 2011) and the Mothur pipeline software tool
(Schloss et al., 2009).
In short, sequences were analysed and sequencing errors were reduced using flowgrams. To reduce
computing time a max number of flows were set to 650-900 depending on the number of available sequences
(50,000-200,000). Subsequently, sequences were trimmed (arbitrary choices: only sequences with min.
lengths of 200 bp and with both primer sequences were selected, tdiffs: 3, and max homop: 8). Sequences
were identified against the Silva reference file release 111 (www.arb-silva.de). To display sequence
abundance a taxonomic tree was calculated with a minimal percentage of all observation of >1 %.
5.3.2 qPCR
For the qPCR measurements employed to determine Leptothrix sp., a Light cycler 480 II of Roche was used.
Samples were taken from both, ‘BW 1st RSF’ and ‘BW A1’. 100mL water was filtered through a 0.45 μm
membrane filter. The filter was used in the ‘powerbiofilm DNA extraction kit’ from MoBio (article number:
24000-50). DNA extraction is based on mechanical and chemical lysis. DNA binds to a silica membrane,
followed by wash steps. After that, DNA is eluted in 100 μl elution buffer from the MoBio kit. Primers and
probe for Leptothrix sp. were heterogeneous for its 16S rRNA. Two upstream primers for Leptothirx sp. are
required to get all relevant species.
Forward primer1 PS-1: 5’ ACGGTAGAGGAGCAATC 3’
Forward primer2 PSP-6: 5’ CAGTAGTGGGGGATAGCC 3’
Reverse primer DSP-6: 5’ GCTTTTGTCAGGGAAGAAATC 3’
Lepto-pr6 forward: 5’ CACGCGGCATGGCT 3’ *Cy5
(Burger et al., 2008a)
(developed by WLN)
The PCR program was as follows: 10 minutes room temperature (uracil glycosylase), 5 minutes 95 °C
(denaturation), 50x 30 seconds at 95 °C; 1 minute at 55 °C and 10 seconds at 72 °C. To quantify the number
of bacteria, a WLN-III plasmid was developed for target genes. The start concentration of this plasmid is
referenced to the mip gene of the standardized Minerva plasmid, and is about 160,000 cDNA/L for
Leptothrix sp.
To determine the gene copy numbers of Pseudomonas spp. and Pseudomonas putida, a quantitative PCR (qPCR)
protocol, using newly developed primers was used. For the specific detection of Pseudomonas spp. primers
were developed targeting the 16S rRNA gene, for detection of P. putida, the more variable gyrB gene was
selected. The primers and probe used for detection of Pseudomonas spp. were Pspp16Sf1 (5’-GAG CCT
AGG TCG GAT TAG-3’), Psppr3 (5’-CGC TAC ACA GGA AAT TCC AC-3’), and probe PsppP1 (5’CGC GTG TGT GAA GAA GGT CTT CG-3’). For quantification of P. putida the primers PpgyrBf3 (5’GAC ATC CTG GCC AAG CGT-3’) and PpgyrBr3 (5’-CTT CCT GYT CGA TGT AGC-3’) and probe
PpgyrBp1 (5’-CTG CAR TGG AAY GAC AGC TTC AAC G-3’) were selected. Primer specificity and
selectivity was analysed and PCR conditions were optimized.
PCR was conducted in 50 µl reaction volumes containing 25 µl of 2x IQ Supermix (Bio-Rad Laboratories
BV, The Netherlands), 10 pmol of the forward, reverse primer and probe, 20 mg of bovine serum albumin,
and 10 µl of the DNA template. The amplification program consisted of 2 min at 95 °C; 43 cycles of 20 s
at 95 °C and 30 sec at 60 °C. Amplification, detection, and data analysis were performed in an iCycler IQ
real-time detection system (Bio-Rad Laboratories BV, The Netherlands). The PCR cycle after which the
fluorescence signal of the amplified DNA and the probe was detected (threshold cycle [Cq]) was used to
quantify the gene copy concentration. Quantification was based on comparison of the sample Cq value with
77
the Cq values of a calibration curve based on known copy numbers of the plasmin containing the 16S rRNA
gene of P. putida (U70977.1) and the gyrB gene of P. putida (HF545867.1).
5.3.3 Matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry (MALDI-TOF MS)
With this technique microorganisms can be identified directly after culturing on selective agar media. Spectra
were generated with the MALDI-TOF MS biotyper from Brϋker Daltonik GmbH, and compared with
approximately 4000 spectra in the Brϋker Daltonik GmbH database. In a log score from 1 to 3, the MALDITOF biotyper defined the similarity of the known and unknown spectra. When the log score is between
2 and 2.3, the genus identification is secure and probable also the species is identified. With a log score
> 2.3, it is highly probable that the species is identified. MALDI-TOF MS is based on the chemotaxonomy
of microorganisms. This ‘fingerprint’ is based on identified proteins of the microorganism. These proteins
are always present in a living cell and make it possible to characterize microorganisms. In this project we
expected Pseudomonas to grow in the filter column where Mn2+ was oxidized (‘BW A1’). Samples, were
filtered through a 0.45 µm membrane and incubated on Pseudomonas specific agar, a media containing
cetrimide and sodium nalidixate to inhibit gram positive bacteria and some gram negatives other than
Pseudomonas.
Pseudomonas secretes a variety of pigments, including pyocyanin (blue-green), pyoverdine (yellow-green and
fluorescent), and pyorubin (red-brown). Coloured colonies on a Pseudomonas agar are suspected to be
Pseudomonas, and they were identified using the MALDI-TOF MS biotyper from Bruker. A single colony of
a target organism is put directly on a 96 target plate. After deposition, the spots were overlaid with 1 µl
matrix solution (2.5mg α-Cyano-4-hydroxycinnamic solved in 50% acetonitrile, 2.5 % trifluoro acetic acid,
47.5 % ultra-pure water). The matrix opens the cell wall. A laser irradiate the matrix sample, to divide it in
little portions of proteins. The matrix evaporate and positive charged proteins become free. In the strong
electric field, the positive charged proteins are lined up. So these proteins have the same starting point,
before they accelerate in the flight tube to get to their specific time-of-flight corresponding with their specific
mass.
5.3.4 Fermenter growth test with selected bacteria to produce
biological MnOx
A Bioflow III fermenter from New Brunswick scientific was inoculated with pure cultures of Pseudomonas
putida (ATCC 23483, LMG 2321), as well as with P. grimontii and P. koreensis, which were isolated from
‘BW A1’. The growth medium used in the fermenter, is described by Jiang et al. (2010). Growth and
subsequent manganese oxidation in the fermenter was performed for 4 days. Growth of the Pseudomonas
species was performed by use of the standard Pseudomonas agar growth medium (48.4 g agar and 10 ml
Glycerol per liter and sterilized for 15 min at 121 °C). The incubation time for growth was 24-48 hrs. at
30 ⁰C. Colonies were identified by MALDI-TOF MS (section 5.3.3) and stored in Pentone-glycerol at 80 ⁰C. Formation of MnOx was identified as black deposits and was verified by ICP-MS and SEM-EDX.
78
5.4.1 Next generation DNA sequencing
Table 5.1 provides an overview of the bacteria population found in backwash water samples from the first
stage filter (‘BW 1st RSF’) and the pilot filter column (‘BW A1’). The sequencing results of sample
‘BW 1st RSF’ are based on 188,241 sequences, whereas the results of sample ’BW A1’ are based on 55,298
sequences. The taxonomic trees are shown in Annex A and B.
Table 5.1: Bacteria population in samples of ‘BW 1 s t RSF’ and ‘BW A1’.
Sample
BW 1st RSF
BW A1
Identification
Gallionella sp.
Other
Nitrospira sp.
Pseudomonas sp.
Gallionella sp.
Other
% of population
97.0
3.0
25,7
14,3
12,4
47,6
Table 5.1 shows a clear difference in bacterial population present in sample ‘BW 1st RSF’ (iron removal filter
of the full scale plant) in comparison to bacteria present in ‘BW A1’ of the freshly ripened (manganese
removal) pilot filter column.
The majority (97 %) of bacteria found in ‘BW 1st RSF’ consisted of Gallionella sp. The abundance of
Gallionella sp. is understandable as iron removal takes place by a biological removal mechanism. Very similar
bacteria composition, is present in the feed water to the 2nd stage full scale and thus also in the feed water
of the pilot filter column. Identification of bacteria present in ‘BW A1’, showed that only 12.4 % of the
bacteria population was of a Gallionella species. Pseudomonas sp. and Nitrospira sp. represented 14.3 % and
25.7 % of the total population, respectively. Almost half of the population found in ‘BW A1’ belongs to
smaller populations or could not be identified. The presence of the bacterium Nitrospira sp. (25.7 %) was
expected, because this species is involved in the oxidation of ammonia and specifically conversion of nitrite
to nitrate, which takes place in this filter. Pseudomonas sp. might be involved in the biological manganese
removal process. Literature suggests that besides oxidation of manganese, Pseudomonas sp. is able to oxidize
ammonia (Daum et al., 1998, Nemergut and Schmidt, 2002). This finding also explains the strong relation
between manganese and ammonia oxidation observed in practice (Bruins et al., 2014). The observation that
14,3 % of the bacteria population found in ‘BW A1’ consists of Pseudomonas sp., a potential manganese
oxidizing bacterium, supports the assumption that manganese removal starts biologically (Bruins et al.,
2015b).
5.4.2 qPCR
Leptothrix sp. and Pseudomonas sp., are both able to oxidize dissolved manganese. Their concentration was
quantified with qPCR as the number of DNA copies (n [cDNA/L]) present in ‘BW A1’. Furthermore the
concentration of the species Pseudomons putida was quantified with the same technique. Table 5.2 gives an
overview of the number of quantified species, expressed as DNA copies/L.
79
Table 5.2: Manganese oxidizing bacteria in sample ‘BW A1’, quantified by qPCR.
Bacterium
Pseudomonas spp.
Pseudomonas putida
Leptothrix sp.
n (cDNA/L)
2.3 x 1011
1.5 * 107
3.8 x 106
%
> 99.99 %
< 0.01 %
< 0.01 %
From table 5.2 it is clear that from the potential manganese oxidizers the presence of Pseudomonas spp. was
much more pronounced, than the presence of Leptothrix sp. This supports the fact that Leptothrix sp. was
not found with the Next generation DNA sequencing (section 3.1). Also Pseudomonas putida was present in
relatively low concentrations, compared to the genus Pseudomonas. In literature Pseudomonas putida is often
associated with biological manganese removal. However, the species Pseudomonas putida was not ubiquitous
and thus not likely responsible for the fast ripening of manganese removal filters in the pilot testing
performed in this study. At the same time it is plausible that other closely related Pseudomonas species
contributed to this process.
5.4.3 Matrix-assisted laser desorption/ionization time-of-flight mass
spectrometry biotyper (MALDI-TOF MS)
Several colonies isolated from sample ‘BW A1’, were cultured with a pseudomonas agar and were identified by
using the MALDI-TOF MS biotyper. Table. 5.3 shows an overview of all identified Pseudomonas species
present in the backwash water of a freshly ripened manganese removal filter.
Table 5.3: Pseudomonas species identified by MALDI-TOF MS (including Log score).
Identified species
Pseudomonas gessardii
Pseudomonas (libanensis)
Pseudomonas (synxantha)
Pseudomonas (veronii)
Pseudomonas (grimontii) Fig. 5.2
Pseudomonas (koreensis)
Pseudomonas (extremorientalis)
Pseudomonas (marginalis)
Pseudomonas (tolaasii)
Pseudomonas (azotoformans)
Pseudomonas (rhodesiae)
Log score
2.40
2.22
2.21
2.17
2.15
2.14
2.07
2.04
2.03
2.03
2.00
The genus Pseudomonas consists of many very closely related species. No Pseudomonas putida was identified in
any of the samples with MALDI-TOF MS. This was expected based on the low contribution of the strain
P. putida, to the total bacterial population determined with qPCR, (< 0.01 %). The list of Pseudomonas species
(table 5.3) is incomplete, as only a limited amount of colonies are identified. SEM images (Fig. 5.2) show
two Pseudomonas species, namely Pseudomonas grimontii and Pseudomonas koreensis in the backwash water of a
freshly ripened manganese removal filter.
80
Figure 5.2: SEM - images (10.000x) of Pseudomonas grimontii (L) and Pseudomonas koreensis (R),
isolated from sample ‘BW A1’.
5.4.4 Fermenter growth test with selected bacteria for the biological
production of MnOx
Pseudomonas grimontii (log: 2.15) and Pseudomonas koreensis (log: 2.14), obtained from ‘BW A1’ , were used as
inoculum in a fermenter to investigate their growth and MnOx production under controlled laboratory
conditions. As a reference, a similar growth test was conducted with the laboratory species Pseudomonas putida
(ATCC 23483, LMG 2321). Although to a very limited extend, Pseudomonas putida was able to produce MnOx.
Results obtained from the fermenter growth test showed that Pseudomonas grimontii and Pseudomonas koreensis
were not able to oxidize Mn2+ producing MnOx, under the performed laboratory conditions.
Summarizing, results from this study show that the population of bacteria present in the backwash water of
the 1st stage (iron removal) filter and the freshly ripened (manganese removal) pilot filter column,
dramatically changed. Furthermore the presence of Pseudomonas putida was very limited (< 0.01 % of the
potential manganese oxidizing bacteria present). This indicates that the role of Pseudomonas putida, concerning
manganese removal at the location Grobbendonk, is limited. However, related Pseudomonas species, may play
an important role in the process of manganese removal (Table 5.3), taking into account that the Birnessite
(MnOx) produced in the pilot filter column during the ripening period was of biological origin (Bruins et al.,
2015b). It remains, however, unclear if Pseudomonas sp. is the only manganese oxidizing bacterium involved
in the initial Mn2+ oxidation, or that other species form a microbial consortium, together, responsible for
the oxidation of Mn2+. The knowledge that manganese removal in aeration-rapid sand filtration treatment
is initiated biologically, together with insight in the manganese oxidizing bacteria species involved, may
enable typically long ripening times to be reduced by creating conditions favorable for the growth of these
manganese oxidizing species. Therefore, the focus of the follow-up research will be on the inoculation of a
consortium of bacteria, identified in manganese removing filters, to enhance filter media ripening. Also
conditions supporting the fast growth of Mn2+ oxidizing bacteria should be investigated in follow up
research.
81
5.5 Conclusions
From this study it can be concluded that:
•
•
•
•
•
Based on ‘next generation DNA sequencing’ analyses, the population of bacteria present in
backwash water from an iron removal filter (first step filter in a full scale plant), and the freshly
ripened pilot manganese removal filter showed a clear population shift from the iron oxidizing
species Gallionella sp. (97 %) to manganese and nitrite oxidizing species (Pseudomonas sp. (14 %) and
Nitrospira sp. (26 %), respectively). However, it should be noted that 47.6 % of the bacteria
population in the manganese removal filter, is still unknown;
qPCR analysis showed that less than 0.01 % of the genus Pseudomonas present, in freshly ripened
manganese removal pilot filter column was of the Pseudomonas putida species;
Pseudomonas sp. is most likely (one of) the manganese oxidizing bacterium genus that play an
important role in the initial stage of the ripening of the manganese removal filters at the full scale
GWTP Grobbendonk. However, it is, still unclear whether this bacterium genus is solitary
operating or acting as part of a microbial consortium;
Amongst others, P. gessardii, P. grimontii and P. koreensis, closely related Pseudomonas species, were
detected by the MALDI-TOF MS analysis, and are likely involved in the manganese removal
process, possibly as a part of a bacterial consortium;
Selected Pseudomonas species from the ripened filter media column namely Pseudomonas grimontii and
Pseudomonas koreensis were not able to produce MnOx under controlled laboratory conditions,
whereas the reference species Pseudomonas putida was able to do so.
This research was financially and technically supported by WLN and the Dutch water companies Groningen
(WBG) and Drenthe (WMD). The auteurs are grateful to Mrs. Marsha van der Wiel and Mr. Pim Willemse
of WLN for providing qPCR, MaldiTOF MS analysis and performing the growth tests and to Mr. Jelmer
Dijkstra of Wetsus for performing SEM-analysis. Thanks also to water company Pidpa (Belgium), for their
willingness to share the data from their groundwater treatment plant and to give the opportunity to perform
a pilot test at Grobbendonk
82
5.7 References
Adams, L.F., Ghiorse, W.C., 1985. Influence of Manganese on Growth of a Sheathless Strain of Leptothrix
discophora. Aplied and environmental Microbiology, 49.3, 556 –562.
Barger J.R., Tebo, B.M., Bergmann, U., Webb, S.M., Glatzel, P., Chiu, V.Q., Villalobos, M., 2005. Biotic and
abiotic products of Mn (II) oxidation by spores of the marine Bacillus sp. strain SG-1. American
Barger, J.R., Fuller. C.C., Marcu. M.A., Brearly A., Perez De la Rosa M., Webb S.M., Caldwell W.A., 2009.
Boogerd, F.C., De Vrind J.P.M., 1987. Manganese oxidation by Leptothrix discophora. Journal of bacteriology,
169.2, 489-494.
Brouwers, G.J., Vrind de J.P.M, Corstjens P.L.A.M., Cornelis P, Baysse C, Vrind de Jong E.W., 1999. CumA,
aGene Encoding a multicopper oxidase, is involved in Mn2+oxidation in Pseudomonas putida GB-1.
Applied and environmental microbiology , 65:4, 1762-1768.
Brouwers G.J., Vijgenboom E., Corstjens P.L.A.M., de Vrind J.P.M., de Vrind-de Jong E.W., 2000. Bacterial
Mn2+ oxidizing systems and multicopper oxidases: an overview of Mechanisms and Functions.
Geomicrobiology Journal, 17.1, 1-24.
Bruins, J.H., Vries, D., Petrusevski, B., Slokar, Y.M., Kennedy, M.D., 2014. Assessment of manganese
removal from over 100 groundwater treatment plants. Journal of Water Supply: Research and
Technology-AQUA, 63.4, 268-280.
Bruins, J.H., Petrusevski, B., Slokar, Y.M., Kruithof, J.C., Kennedy, M.D., 2015a. on line. Manganese
removal from groundwater: characterization of filter media coating. Desalination & Water Treatment,
in press, doi: 10.1080/19433994.2014.927802.
Bruins, J.H., Petrusevski, B., Slokar Y.M., Huysman, K., Joris, K., Kriuithof J.C., Kennedy, M.D., 2015b.
Biological and physicochemical formation of Birnessite during the ripening of manganese removal
filters. Water research 69, 154-161.
Burger, M.S., Krentz, C.A., Mercer, S.S., Gagnon, G.A., 2008a. Manganese removal and occurrence of
Burger, M.S., Mercer, S.S., Shupe, G.D., Gagnon, G.A., 2008b. Manganese removal during bench-scale
biofiltration. Water Research 42, 4733-4742.
Caspi, R., Tebo, B.M., Haygood, M.G., 1998. C-Type Cytochromes and Manganese Oxidation in Pseudomonas
putida MnB1. Applied and environmental Microbiology, 64.10, 3549-3555.
Corstjens, P.L.A.M., de Vrind, J.P.M., Goosen, T., de Vrind-de Jong, E.W., 1997. Identification and
molecular analysis of the Leptothrix discophora SS-1 mofA gene, a gene putatively encoding a
manganese-oxidizing protein with copper domains. Geomicrobiology Journal, 14.2, 91-108.
83
DePalma, S.R., 1993. Manganese oxidation by Pseudomonas putida, PhD-thesis, Harvard University,
Cambridge, USA.
De Schamphelaire, L. , Rabaey, K. , Boon, N. , Verstraete, W. and Boeckx, P., 2007. Minireview: The
potential of enhanced manganese redox cycling for sediment oxidation. Geomicrobiology Journal,
24: 7, 547-558.
Daum, M., Zimmer, W., Papen, H., Kloos, K., Nawrath, K., Bothe, H., 1998. Physiological and Molecular
Biological Characterization of Ammonia Oxidation of the Heterotrophic Nitrifier Pseudomonas putida.
Current Microbiology, 37, 281–288.
El Gheriany I.A., Bocioaga B., Anthony Hay A.D., Ghiorse W.C., Shuler M.L., Lion L.W., 2009. Iron
Requirement for Mn (II) Oxidation by Leptothrix discophora SS-1. Applied and Environmental
Microbiology, 75.5, 1229-1235.
Farnsworth, C.E., Voegelin, A. and Hering, J.G., 2012. Manganese oxidation induced by water table
fluctuations in a sand column. Environ. Sci. Technol. 46, 277-284.
Feng, A., Tan, W., Liu, F. Huang, Q., and Liu, X., 2005. Pathways of Birnessite formation in alkali medium.
Sci. China Ser. D., 48.9, 1438-1451.
Fleming, E.J., Langdon, A.E., Martinez-Garcia, M., Stepanauskas, N.J., Poulton, Masland, E.D.P., Emerson,
D., 2011. What’s New Is Old: Resolving the Identity of Leptothrix ochracea Using Single Cell
Genomics, Pyrosequencing and FISH. PLoS ONE, 6.3, 1-10.
Forrez, I., Carballa, M., Verbeken, K., Vanhaecke, L., Schlüsener, M., Ternes, T., Boon, N., Verstrete, W.,
2010. Diclofenac oxidation by biogemic manganese oxides. Environmental science & technology,
44.9, 3449-3454.
Geszvain, K., McCarthy, J.K., Tebo, B.M., 2013. Elimination of Manganese (II, III) Oxidation in Pseudomonas
putida GB-1 by a Double Knockout of Two Putative Multicopper Oxidase Genes. Applied and
Environmental Microbiology, 79 .1, 357–366.
Gounot A-M., 1994. Microbial oxidation and reduction of manganese: Consequences in groundwater and
applications. FEMS Microbiology reviews 14, 339-350.
Hope C.K., Bott T.R., 2004. Laboratory modelling of manganese biofiltration using biofilms of Leptothrix
doscophora. Water research 38, 1853-1861.
Jiang, S., Kim, D-G., Kim, J., Ko, S-O., 2010. Characterization of the biogenic manganese oxides produced
by Pseudomona putida strain MnB1. Environ. Eng. Res., 15.4, 183-190.
Kim, S.S., Bargar, J.R., Nealson, K.H., Flood, B.E., Kirschvink, J.L., Raub, T.D., Tebo, B.M., Villalobos,
M., 2011. Searching for Biosignatures Using Electron Paramagnetic Resonance (EPR) Analysis of
Manganese Oxides. Astrobiology, 11.8, 775-786.
84
Mann, S., Sparks. N.H.C., Scoot G.H.E., Vrind- De Jong E.W., 1988. Oxidation of Manganese and
formation of Mn3O4 (Hausmannite) by spore Coats of a marine Bacillus sp. Applied and
environmental microbiology, 54.8, 2140-2143.
Meerburg, F., Hennebel, T., Vanhaecke, L., Verstraete, W., Boon, N., 2012. Diclofenac and 2anilinophenylacetate degradation by combined activity of biogenic manganese oxides in silver.
Microbial technology, 5.3, 388-398.
Nemergut, D.R., Schmidt, S.K., 2002. Disruption of narH, narJ and moaE inhibits heterotrophic nitrification
in Pseudomonas strain M19. Applied and environmental biology, 68.12, 6462-6465.
Post, J.E. and Veblen, D.R., 1990. Crystal structure determinations of synthetic sodium, magnesium, and
potassium birnessite using TEM and the Rietveld method. American mineralogist, 75, 477-489.
Post, J. E., 1999. Manganese oxide minerals: Crystal structures and economic and environmental
significance. Proceedings of the National Academy of Sciences USA, 96, 3447-3454.
Sabirova, J.S., Cloetens, L.F.F., Vanhaecke, L., Forrez, I., Verstraete, W., Boon, N., 2008. Manganeseoxidizing bacteria mediate the degradation of 17α-ethinylestradiol. Microbial Biotechnology, 1.6, 507512.
Schloss, P.D. Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakly,
B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B., Thallinger, G.G., Van Horn, D.J., Weber, C.J.,
2009. Introducing Mothur: Open source, platform-independent, community-supported software for
describing and comparing microbial communities. Appl. Environ Microbiol, 75 (23), 7537-7541.
Schloss, P.D., Gevers, D., Westcott, S.L., 2011. Reducing the effects of PCR amplification and sequencing
artifacts on 16S rRNA-based studies. PloS One. 6 (12): e27310, doi:10.1371/journal.pone.0027310.
Stumm, W. Morgan, J.J., 1996. Aquatic Chemistry, chemical equilibria and rates, 3rd ed. Wiley New York.
Tebo, B.M., Marger J.R., Clement B.G., Dick G.J., Murray K.J., Parker. D., Verity R., Webb S.M., 2004.
32, 287-328.
Tebo, B.M., Johnson H.A., McCarthy J.K., Templeton A.S., 2005. Geomicrobiology of manganese (II)
Oxidation. TRENDS in Microbiology, 13.9, 421-428.
Tekerlekopoulou, A.G., Vasiliadou I.A., Vayenas D.V., 2008. Biological manganese removal from potable
water using trickling filters. Biochemical Engineering Journal 38, 292-301.
Vandenabeele, J., de Beer, D., R. Germonpré, R., Verstreate, W., 1992. Manganese oxidation by microbial
consortia from sand filters. Microbial Ecology 24, 91-108.
Vandenabeele J., 1993. Manganese removal by microbial consortia from rapid sand filters treating water
containing Mn2+ and NH4+, PhD thesis, Ghent, Belgium.
Verstraete, W., 2013, Personal communication, Laboratory of microbial ecology and technology,
department of biochemical and microbial technology, Ghent University, Belgium.
85
Villalobos, M., Toner, B., Bargar, J., Sposito, G., 2003. Characterization of the manganese oxide produced
by Pseudomonas putida strain MnB1. Geochimica et Cosmochimica Acta, 67.14, 2649-2662.
Villalobos, M., Lanson, B., Manceau, A., Toner, B., Sposito, G., 2006. Structural model for the biogenic Mn
Vrind de, J.P.M., Boogerd, F.C., Vrind de – Jong de, E.W., 1986. Manganese reduction by a marine Bacillus
Species. Journal of Bacteriology, 167.1, 30-34.
Webb, S.M., Tebo, B.M., Bargar, J.R., 2005a. Structural characterization of biogenic Mn oxides produced in
seawater by marine bacillus sp. strain SG-1. American mineralogist, 90, 1342-1357.
Webb, S.M., Dick, G.J., Bargar, J.R., Tebo, B.M., 2005b. Evidence for the presence of Mn (III) intermediates
in the bacterial oxidation of Mn (II). Proceedings of national academic science, 102.15, 5558-5563.
86
Annex A
Next generation DNA sequencing, taxonomic trees of sample ‘BW
1st RSF’ (based on 188,241 sequences)
Root, 100.01115716%
Bacteria, 100.01115716%
Firmicutes, 1.176283073%
Clostridia, 1.1672510892%
Clostridiales, 1.166719796%
Proteobacteria, 98.549569653%
Betaproteobacteria, 97.590054192%
Nitrosomonadales, 97.327595367%
Gallionellaceae, 97.324407608%
Gallionella, 97.020507916%
87
Annex B
Next generation DNA sequencing, taxonomic trees of sample
‘BW A1’ (based on 55,298 sequences)
Root, 100.25926933%
Bacteria, 100.25926933%
Actinobacteria, 1.4975976793%
Actinobacteria, 1.4975976793%
Nitrospirae, 25.785513553%
Nitrospira, 25.785513553%
Nitrospirales, 25.785513553%
Nitrospiraceae, 25.76738283%
Nitrospira, 25.656785423%
Planctomycetes, 1.0715257003%
Planctomycetacia, 1.0207596773%
Planctomycetales, 1.0207596773%
Planctomycetaceae, 1.0207596773%
Proteobacteria, 69.284742997%
Alphaproteobacteria, 15.85169069%
Rhizobiales, 3.8781615447%
Methylocystaceae, 1.6027558698%
uncultured, 1.5973166531%
Nordella, 1.0588341945%
Sphingomonadales, 11.286374762%
GOBB3-C201, 5.6386547004%
Sphingomonadaceae, 3.374127459%
Blastomonas, 1.6825310489%
Novosphingobium, 1.0298250385%
Betaproteobacteria, 29.185024023%
Burkholderiales, 14.060375306%
Comamonadaceae, 9.3046867918%
Rhodoferax, 2.8138881334%
Oxalobacteraceae, 4.7212401414%
Undibacterium, 3.7711902819%
Nitrosomonadales, 13.97516091%
Gallionellaceae, 12.856495331%
Gallionella, 12.39597498%
Nitrosomonadaceae, 1.1186655788%
Gammaproteobacteria, 23.301604569%
Methylococcales, 1.7822500227%
Methylococcaceae, 1.7568670112%
Methylobacter, 1.2945335872%
Pseudomonadales, 19.229444293%
Pseudomonadaceae, 18.859577554%
Azomonas, 4.4891668933%
Pseudomonas, 14.32145771%
88
89
Figure: (top) GWTP Grobbendonk (2015, Koen Joris, Pidpa) and (bottom) GWTP De Punt (2014,
WBG)
90
6
REDUCTION OF RIPENING TIME OF FULL SCALE
MANGANESE REMOVAL FILTERS WITH MANGANESE
OXIDE COATED MEDIA
Kruithof, Maria D. Kennedy (2015). Reduction of ripening time of full scale manganese removal filters
with manganese oxide coated media. Journal of Water Supply: Research and Technology – AQUA
64.4: 434 – 441
91
6.1 Abstract
Effective manganese removal by conventional aeration-filtration with virgin filter media requires a long
ripening time. The aim of this study was to assess the potential of manganese oxide-coated media to shorten
the ripening time of filters with virgin media, under practical conditions. A full scale filter filled with virgin
sand and a full scale filter filled with anthracite/sand were operated at two groundwater treatment plants, in
parallel with (full scale) test filters, with an additional layer of Manganese Oxide-Coated Sand (MOCS) or
Manganese Oxide-Coated Anthracite (MOCA). Significantly different ripening times were observed to
achieve an effective manganese removal: 55 days for a filter with virgin sand and 16 days for a filter with
virgin anthracite/sand, respectively. The observed differences could be attributed to different feed water
quality, different iron loading, and backwashing intensity and frequency. In batch experiments fresh MOCA
and MOCS showed good manganese adsorptive properties. Addition of a shallow layer of fresh MOCA in
test filters eliminated the ripening time, while a layer of aged MOCS did not significantly shorten the ripening
period. The poor performance of the aged MOCS was likely caused by changed properties of aged and dried
MOCS that had lost its adsorption capacity, the auto-catalytic activity and the biological activity.
Keywords: filter media, filter ripening, groundwater treatment, manganese oxide coating, manganese
removal
6.2 Introduction
In some countries (e.g., the US, and Central and Eastern Europe) an efficient manganese removal is
commonly achieved by pre-oxidation with strong oxidants, such as O3, Cl2, ClO2, KMnO4, followed by
rapid sand filtration. Use of strong oxidants for manganese removal is not desirable due to the potential
formation of harmful oxidation by-products, as well as costs and risks associated with the usage and handling
of chemicals. In some cases pre-oxidation is combined with filtration through a filter bed with a manganese
adsorbent, most frequently manganese green sand. This treatment can be very effective but it requires
continuous or intermittent regeneration typically with potassium permanganate (Knocke et al., 1991). It is
in view of the above mentioned disadvantages, that the removal of manganese from groundwater in The
Netherlands and Belgium is normally achieved with conventional aeration-filtration treatment, also called
contact filtration. Under common groundwater conditions (e.g., low pH), manganese removal may be
initiated by bacterial activity during aeration-filtration (Diem & Stumm, 1984; Burger et al., 2008). Although
aeration-filtration is efficient and cost effective, it is in practice, however, frequently associated with a major
drawback:
•
very long ripening times of virgin filter media; several weeks to more than a year (Fig. 6.1), are
required to achieve an efficient manganese removal (Buamah et. al., 2009a; Cools, 2010; Krull,
2010).
Ripening of filter media (for manganese removal), is defined as the development of properties to autocatalytically adsorb and subsequently oxidise Mn2+, without the use of strong oxidants, such as Cl2, O3 and
KMnO4.
92
100
80
60
40
20
0
0
40
80
120
160
200
240
280
320
360
Figure 6.1: An example of a very long ripening time of virgin filter media for manganese removal in a
Dutch full scale groundwater treatment plant (GWTP Baanhoek, Evides Water Company).
From Fig. 6.1, it can be seen that efficient manganese removal in a filter with virgin sand was not achieved
until after almost one year of continuous filter operation. However, such a long ripening time is exceptional,
typically it takes 1 to 4 months to achieve an efficient manganese removal.
Many authors (Hu et al., 2004a; Kim & Jung, 2008; Kim et al., 2009) describe the potential of Manganese
Oxide-Coated Sand (MOCS) to adsorb dissolved manganese from (ground)water. It was also reported that
removal of Mn2+ in filters with anthracite is enhanced by development of ‘catalytic oxide layers’ on aged
anthracite, due to formation of Manganese Oxide-Coated Anthracite (MOCA) (Sahabi et al., 2009). Buamah
et al. (2008) suggested that the performance of conventional manganese removal plants could be improved
by introducing manganese and/or iron (hydro)oxide-rich filter media into rapid sand filters.
The primary aim of the study reported in this paper was to examine if the long ripening time typically
required to achieve an effective manganese removal with virgin filter media, in full scale conventional
aeration-filtration treatment plants, could be substantially reduced by addition of a MOCS or MOCA layer.
6.3.1 MOCS and MOCA
The manganese oxide-coated filter media used in this research were obtained from two full scale
groundwater treatment plants (GWTPs). MOCS was obtained from GWTP De Punt (Water Supply
Company Groningen, The Netherlands), and MOCA was obtained from GWTP Grobbendonk (Pidpa
water supply company, Belgium). It was shown that both MOCS and MOCA coatings contain a Birnessite
type of manganese oxide (Bruins et al., 2014a).
For all batch adsorption experiments and full scale filter experiments fresh MOCA or MOCS were taken
directly from an operating, ripened full scale manganese removal filter. However, MOCS was stored in the
open air for several months prior to the full scale filter experiments.
93
6.3.2 Physical and chemical properties of MOCS and MOCA
Chemical composition of the MOCS and MOCA coating was determined by boiling the media in 3 M
HNO3, followed by analysis for Fe, Mn, Ca, Si and Al with inductively coupled plasma mass spectrometry
(ICP-MS) according to NEN-EN-ISO 17294-2 (NEN 2004).
The pH of point of zero charge (pHPZC), i.e., the surface charge of coated filter media depending on structural
deficits, unbalanced bonds and the presence of protons, was determined by a mass titration method (Fiol
& Vilaescusa, 2009).
6.3.3 Batch adsorption experiments
To determine the MOCS and MOCA manganese adsorption capacity, batch adsorption isotherm
experiments were carried out. Model water used in these experiments contained 1 mmol/L HCO3- and
2 mg/L Mn2+ in demineralised water; the pH was adjusted to 7 with 0.1 M HCl. The bottles containing
model water and five different concentrations of either MOCS or MOCA in granular form were agitated on
an Innova 2100 shaker at 100 rpm for 48 h. Prior to measuring the final concentration of manganese, the
samples were filtered through a 0.45 μm membrane filter, and acidified to preserve them for the analyses.
The manganese concentration was measured, and results were plotted as a Freundlich adsorption isotherm.
The results obtained from the batch adsorption experiments of the two manganese-coated filter media were
compared with those obtained with a commercial manganese adsorbent Aquamandix (Aqua-techniek, The
Netherlands).
6.3.4 Full scale filter runs
Full scale filter runs were conducted at GWTPs where the manganese-coated filter media were obtained. In
total, six different combinations of filter media were used (Fig. 6.2). Two filters were operated at GWTP
De Punt (Fig. 6.2A-B), and the other four (Fig. 6.2C-F) at GWTP Grobbendonk.
94
Figure 6.2: Schematic presentation of filter media layers in the six full scale filters included in the
study (all values in cm).
Of the two filters at GWTP De Punt, one filter (Fig. 6.2A) was filled with virgin quartz sand, commonly
applied at this plant. This filter served as a reference filter. The second filter (Fig. 6.2B) was filled with the
same virgin sand, however, a 15 cm layer of sand was replaced by MOCS. The MOCS was aged and dried
prior to use in the full scale filter run. Based on practical experience, the MOCS layer was placed at the level
where manganese removal in ripened filters is observed at this facility.
The effect of the MOCA layer on the ripening of virgin filter media was studied at GWTP Grobbendonk.
This plant utilises filters with dual media - anthracite and sand - which is how the reference filter was
prepared (Fig. 6.2C). In the other three (full scale) test filters (Fig. 6.2D-F), part of the top anthracite layer was
replaced by 10, 20 and 30 cm MOCA respectively. MOCA was placed on the top of the anthracite layer,
because at this level manganese removal is observed at this water treatment plant.
The composition of the feed water for the full scale experiments at the two GWTPs is given in Table 6.1.
At GWTP De Punt, groundwater was aerated prior to the test filters. At GWTP Grobbendonk the treatment
consists of a 1st stage rapid sand filtration (aeration and biological adsorptive iron removal), a pH correction
with milk of lime and 2nd stage dual media filtration. The feed water for the experiments at GWTP
Grobbendonk was the water after pH correction.
95
Table 6.1: Composition of feed water during the experiments with full scale filters.
Parameter
Unit
De Punt
Grobbendonk
Iron
mg/L
4.5 - 6.9
0.03 - 0.14
Manganese
mg/L
0.18 - 0.25
0.12 - 0.18
Ammonium
mg/L
0.29 - 0.78
< 0.05 - 0.23
pH
[-]
7.3 -7.5
7.5 - 7.6
Oxygen
mg/L
8 - 10
> 10
Redox potential
mV
-50 to +50
+ 200 to +300
From Table 6.1 it is evident that the feed water quality at the two test locations differed significantly. In
particular the difference in water quality parameters that are known to influence manganese removal (Fe2+,
NH4+ concentrations, pH and redox potential) should be noted (Bruins et al., 2014b).
Table 6.2: Process design parameters and operational conditions of full scale manganese removal
experiments at GWTP De Punt and GWTP Grobbendonk.
Parameter / condition
Unit
De Punt
Grobbendonk
Type of aeration
spray
cascade
Position of filter
‘pre-filter’
‘post-filter’
Type of filtration
down flow
down flow
Type of filter media
quartz sand
anthracite / quartz
sand
Grain size fraction virgin media
mm
1.8 - 2.4
0.8-1.8 / 0.4-0.8
Filter area
Filter bed height
Flow per filter
Filtration rate
Empty bed contact time
Backwash (BW) criterion
Backwash frequency
Filter bed expansion during BW
m2
m
m3/h
m3/m2.h
min
n/week
-
12.5
2
60
4.8
25
head loss
2
no
37.5
1.1 (0.6+0.5)
190
5.0
13.2
head loss
0.5
yes (anthracite)
Filtered volume between BW
Iron loading per filter run (FR)
m3 per filter run
kg Fe/m2.FR
5,000 – 7,000
2.5
10,000
< 0.1
Table 6.2 depicts an overview of process design parameters and operational conditions applied during the
test filter runs at both locations. From this table it is evident that especially ‘Iron loading per filter run’ is
substantially different (De Punt: 2.5 kg Fe/m2.FR and Grobbendonk: < 0.1 kg Fe/m2.FR). Iron loading is
known to influence manganese removal (Bruins et al., 2014b). In Table 6.3 backwash procedures at both
GWTPs are listed.
96
Table 6.3: Backwash procedures applied at full scale plants De Punt and Grobbendonk.
GWTP De Punt
Vf (m3/m2.h)
Duration
Water
Air
Minutes
24
1
52
5
13.6
28
3
24
28
5
24
8
13.6
3
GWTP Grobbendonk
Vf (m3/m2.h)
Duration
Water
Air
Minutes
10.6
0.33
60
1
23.8
5
-
6.4 Results and discussion
6.4.1 Ripening of virgin filter media in reference filters
In Fig. 6.3 the ripening times of the two full scale reference filters filled with virgin sand (GWTP De Punt)
and anthracite/sand (GWTP Grobbendonk) are shown.
100
80
60
40
20
0
0
10
20
30
40
50
60
reference anthracite/sand (Grobbendonk)
reference sand (De Punt)
Figure 6.3: Comparison of the Mn removal efficiency (%) as a function of the filter ripening time for
the two reference full scale filters: (virgin) anthracite/sand, at GWTP Grobbendonk (pH: 7.5 to7.6;
redox potential: +200 to +300 mV; filtration rate: 5.0 m3/m2.h) and (virgin) sand, at GWTP De
Punt (pH: 7.3 to 7.5; redox potential: -50 to +50 mV; filtration rate: 4.8 m3/m2.h).
As seen from Figure 6.3, the ripening time required to reach > 90% manganese removal of the single media
reference filter at GWTP De Punt was about 3.5 times longer (55 days) than that of the dual media reference
filter at GWTP Grobbendonk (16 days). The observed difference was attributed to both the different feed
water quality and the difference in applied operational conditions (e.g., backwashing pattern, intensity and
frequency).
97
6.4.2 Effect of water quality parameters
It has been reported that ferrous iron competes with Mn2+ for adsorption sites on filter media (Hu et al.,
2004a, b). Feed water at GWTP De Punt had a 50 to 150 times higher Fe2+ concentration than the GWTP
Grobbendonk feed (Table 6.1). At GWTP De Punt (in combination with slightly lower feed water pH), the
high Fe2+ concentration caused a more pronounced competition for available adsorption sites by Mn2+. It
has been reported that the presence of iron hydro-oxide layers in the filter media coating could support the
Mn2+ adsorption (Buamah, 2009). However iron hydro-oxide has a much lower Mn2+ adsorption capacity
than manganese (hydro-) oxides (Buamah et al., 2008). Therefore formation of iron hydro-oxide layers in
the filter bed zone where manganese is removed should be prevented in practice.
The feed water concentration of Fe2+ also determines an important operational condition for manganese
removal, i.e., iron loading per filter run. Feed water at GWTP De Punt contained a significantly higher Fe2+
concentration, resulting in much higher iron loading per filter run than at GWTP Grobbendonk (2.5 and <
0.1 kg Fe/m2, respectively). In addition, this much higher iron loading at GWTP De Punt required an
approximately 4 times more frequent backwashing than at GWTP Grobbendonk. Backwashing results in a
partial removal of MnOx from the coating, while the presence of MnOx is essential for an effective
manganese removal. Intensive backwashing can also cause a substantial removal of the biological activity
(e.g., by removal of bacteria from filter media), which may play an important role in the process of manganese
adsorption and oxidation (Vandenabeele et al., 1992; Katsoyiannis & Zouboulis, 2004; Tebo et al., 2004).
Partial loss of MnOx and biological activity is even more pronounced for backwashing with combined water
and air flushing. Both the frequency and intensity of the backwashing at GWTP De Punt were more
detrimental for the filter media, resulting in a longer ripening time of the filters at this location.
Another water quality parameter playing an important role in manganese removal is the pH. To achieve an
effective manganese removal the pH should preferably be above 7.1 (Bruins et al., 2014b). In general, the
higher the pH, the better the manganese removal. From the pH values of feed water at both GWTPs it can
be concluded that Grobbendonk water (pH 7.5 - 7.6) provided slightly better conditions for an effective
manganese removal than GWTP De Punt water (pH 7.3 - 7.5).
NH4+ removal efficiency is another parameter that shows a strong positive correlation with manganese
removal (Bruins et al., 2014b). The presence of NO2- (due to incomplete NH4+ removal) may not only
prevent effective Mn2+ removal, but may even cause manganese leaching, by reducing already adsorbed and
oxidised MnOx back to Mn2+ (Vandenabeele et al., 1995). The NH4+ concentration in the filtrate of the two
filters at GWTP De Punt was substantially higher than at GWTP Grobbendonk especially during the filter
start-up, which could explain the much longer ripening time of filters at the De Punt location.
Finally, oxidation of (adsorbed) manganese takes place more easily at a higher redox potential (Stumm &
Morgan, 1996; Scherer & Wichmann, 2000; Flemming et al., 2004). The redox potential of the feed water at
GWTP De Punt was much lower compared to that of the feed water at GWTP Grobbendonk (50 / + 50 mV and +200 / +300 mV, respectively), partly caused by the presence of NH4+ in the filtrate,
which was much higher at the De Punt location. As a consequence, even if Mn2+ was adsorbed it was not
as effectively oxidised and therefore possibly desorbed, resulting in a longer ripening time at the De Punt
location.
Although the feed water quality at both locations is suitable to achieve effective manganese removal in a
conventional aeration-filtration system once the filters are ripened, the conditions to achieve shorter filter
ripening times were found to be more favourable at the Grobbendonk location.
98
In summary, much faster ripening of virgin filter media with respect to complete manganese removal at
GWTP Grobbendonk can be attributed to the combined effect of the following parameters crucial for
manganese removal:
• more favourable feed water quality (lower Fe2+ and NH4+ concentrations, higher pH and redox
potential).
• more favourable operational conditions (lower iron loading per filter run, lower backwash
frequency and intensity).
6.4.3 MOCS and MOCA characterisation and batch adsorption
experiments
Table 6.4 depicts the physical characteristics of MOCS, MOCA, and Aquamandix (AQM), and coating
composition of MOCS and MOCA.
Coating
composition
Physical properties
Table 6.4: Physical characteristics of MOCS, MOCA and AQM, and the coating composition of
MOCS and MOCA.
Parameter
Unit
MOCS
MOCA
AQM*
Bulk density
kg/L
1.177
0.650
2.000
Particle density
kg/L
2.326
1.176
3.600
Porosity
%
49.4
44.7
44.4
Grain size (d10-d90)
mm
1.6 - 3.1
0.8 - 1.5
1.0 - 2.0
Uniformity coefficient
[-]
1.58
1.52
pHPZC
[-]
7.2
8.0
5.0
Mn
Fe
Ca
mg/g
mg/g
mg/g
12.8
158
8.85
13.5
2.22
2.45
-
Si
Al
mg/g
mg/g
14.5
0.47
0.28
0.25
-
* According to the supplier, Aquamandix consists of 78% MnO2, 6.2% Fe2O3, 5.2% SiO2, 3.1% Al2O3.
From Table 6.4 it can be seen that the grain size of MOCA is considerably smaller than that of MOCS. As
a consequence, MOCA has a larger geometric surface area, enhancing the adsorption capacity. Another
characteristic indicating a better adsorption capacity of MOCA is the coating composition. The most
pronounced difference in chemical composition of the two media is the iron content. Iron is present as iron
(hydr)oxide, whereas manganese is present as manganese oxide (MnOx). Although both oxides can absorb
Mn2+, iron (hydr)oxide has a significantly lower manganese adsorption capacity (Buamah et al., 2008).
Therefore it was expected that MOCA with a more than 70 times lower iron content, would adsorb Mn2+
better. On the other hand the pHPZC of MOCS was significantly lower compared to MOCA. This suggests
that MOCS will have better adsorptive properties over a wider pH range for positively charged ions such as
Mn2+.
In Table 6.5 the Freundlich adsorption isotherm constants for manganese adsorption on MOCS, MOCA,
and Aquamandix are given.
99
Table 6.5: Freundlich adsorption isotherm constants for Mn 2 + adsorption on MOCS, MOCA and
Aquamandix.
Constant
K [(mg/g) / (mg/L)]
1/n
r2
qe (mg/g)
at Ce=0.2 mg/L
qe (g /L)
at Ce=0.2 mg/L
MOCS
0.45
1.31
0.91
0.132
Adsorbent
MOCA
0.91
1.34
0.91
0.276
Aquamandix
0.90
1.38
0.96
0.280
0.155
0.179
0.560
From Table 6.5 it can be seen that manganese adsorption capacities, qe expressed per unit weight of
adsorbent, for MOCS and MOCA are very different (0.132 and 0.276 mg/g ads., respectively, at Ce of
0.2 mg/L Mn2+). When expressed per unit volume, however, adsorptive capacities of MOCS and MOCA
were found to be similar (0.155 and 0.179 mg Mn2+/L of MOCS and MOCA, respectively, at Ce of 0.2 mg
Mn2+/L). A much higher adsorption capacity per volume of adsorbent was found for Aquamandix.
However this commercial adsorbent has no auto-catalytic oxidation properties, thus once the adsorption
capacity is exhausted, manganese removal stops (Buamah, 2009).
Based on the adsorption capacities, qe, reported in Table 6.5, the (calculated) theoretical manganese
adsorption capacities of the MOCS and MOCA layers, placed in the full scale test filters are calculated Table
6.6.
Table 6.6: Calculated theoretical adsorptive capacity of MOCS and MOCA layers with associated
duration of filter run (FR) time before the manganese breakthrough.
Filter
Filter bed media
Capacity1
FR time before Mn
breakthrough2
(kg)
(hrs)
A
B
C
D
E
F
reference sand
sand + 15 cm MOCS
reference anthracite/sand
anthracite/sand + 10 cm MOCA
0
0.29
0
0.67
1.34
2.01
24 - 48
24 - 48
48 - 96
72 - 144
1 At Ce=0.2 mg/L
2 For Vf between 2.5 - 5 m3/m2.h
This Table (6.6) also shows the expected operational times of the filters before manganese breakthrough,
assuming that adsorption was the only manganese removal mechanism ignoring the catalytic effect
associated with adsorption and oxidation of adsorbed manganese. For a 15 cm MOCS layer (De Punt), the
Mn2+ adsorption capacity is approximately 0.29 kg, with an expected breakthrough after 24 to 48 hrs,
whereas these values for a 10 cm layer of MOCA (Grobbendonk) are respectively 0.67 kg Mn2+ and
breakthrough also after 24 to 48 hrs.
100
6.4.4 Ripening of full scale filters with the addition of MOCS and
MOCA layers
100
M nremoval efficiency (%)
Manganese removal during the ripening time of the two full scale filters at GWTP De Punt are shown in
Fig. 6.4(L). Based on batch adsorption experiments conducted with fresh MOCS (Tables 6.5 and 6.6), it was
expected that the MOCS layer in the full scale filter would effectively remove Mn2+ at least during the first
24 - 48 hrs. However, the addition of 15 cm of MOCS did not have a significant impact on the ripening
time of the filter with respect to the manganese removal. The reason for the poor Mn2+ removal is probably
attributed to the difference in adsorptive properties of MOCS used in the batch adsorption experiments
(freshly taken from a running ripened filter), and in the filter runs (dried and stored for several months
before use). The layered structure of Birnessite may have irreversibly collapsed, decreasing the number of
available adsorptive sites (Post, 1999). Furthermore, long exposure to air may have resulted in (complete)
oxidation of the auto-catalytically active Birnessite into not auto-catalytically active Pyrolusite (MnO2).
Storage of the MOCS could also have resulted in a loss of biological activity (e.g., by dying of bacteria present
on MOCS), which may play an important role in initiating oxidation of manganese adsorbed on filter media
(Vandenabeele et al., 1992; Katsoyiannis & Zouboulis, 2004; Tebo et al., 2004). Drying and storage of the
MOCS used in the full scale filter most likely caused loss of a substantial part of its original adsorption
capacity. Besides, it is most likely the stored MOCS had lost its auto-catalytic and biological activity.
80
60
40
20
0
0
10
20
30
40
50
reference sand
sand + 15 cm MOCS
60
100
80
60
40
20
0
0
10
20
30
40
reference anthracite
10 cm MOCA
20 cm MOCA
30 cm MOCA
Figure 6.4: Mn removal efficiency (%) as a function of the ripening time of two (full scale) test filters
at GWTP De Punt (L) - (pH: 7.3 to 7.5; redox potential: -50 to +50 mV; filtration rate: 4.8
m 3 /m 2 .h) and four full scale test filters at GWTP Grobbendonk (R) - (pH: 7.5 to7.6; redox potential:
+200 to +300 mV; filtration rate: 5.0 m 3 /m 2 .h).
In Fig. 6.4(R) the results of the experiments with four full scale test filters at GWTP Grobbendonk are
shown. At this location the filter ripening time of the reference anthracite/sand filter was compared to the
ripening times of three test filters containing MOCA layers of different thicknesses (Figs. 6.2D-F).
Results obtained with the test filters containing a layer of MOCA showed a very high (> 90%) manganese
removal from the start of the filter run, irrespective of the thickness of the MOCA layer. As a comparison,
the manganese removal efficiency in the reference filter without a MOCA layer was approximately 10%
during the first 10 days of operation.
Assuming that the manganese removal in the test filters was achieved only by adsorption on MOCA, the
filter with a 10 cm layer should display breakthrough after 24 hrs of filter operation (Table 6.6). However,
no breakthrough was observed during more than a month of continuous operation of the test filters, most
likely due to the presence of Birnessite and/or manganese oxidising bacteria in the fresh MOCA, promoting
autocatalytic oxidation, resulting in the immediate formation of a new active MnOx layer.
101
After about 2.5 days and in particular after 15 days of operation, a decrease in manganese removal efficiency
was observed for the test filter with a 10 cm layer of MOCS. The reason was an operational malfunctioning,
caused by a poor distribution of the feed water over the filter surface. Additionally, the feed water jet
disrupted the top of the MOCA layer, locally thinning the layer to less than 10 cm. As soon as the MOCA
layer was restored by a gentle backwash, manganese removal efficiency was re-established.
Based on the results depicted in Fig. 6.4(R), a 10 cm layer of fresh MOCA is sufficient to achieve an efficient
manganese removal from the start of a new filter. However, to prevent practical problems (e.g., short
circuiting), it is recommended to apply a MOCA layer of at least 20 cm thickness.
Comparing the results of the full scale test filters with an added MOCS (Fig. 6.2B) and MOCA layer (Fig.
6.2D-F), a poor ripening of filters containing MOCS and a fast ripening of filters containing MOCA was
observed. The major reason for the poor results achieved with the MOCS layer in a filter was probably, as
explained above, attributed to the use of aged (dried) MOCS. The difference in performance could also be
caused by the different feed water quality (e.g., redox potential, NH4+ removal, pH), difference in MOCS
and MOCA composition (coating Fe content) and different operational conditions applied (grain size of
MOCS is approximately double that of MOCA, and Fe2+ loading and backwashing pattern and frequency
were significantly different).
To summarise, this research showed that fresh manganese oxide-coated filter media were able to shorten
the filter media ripening time substantially. Drying the MOCS has affected the results dramatically. In
addition differences may have been caused by water quality as well as operational conditions. These
phenomena must be investigated in more detail under comparable conditions, emphasising also the role of
micro-biology and the importance and influence of specific bacteria.
6.5 Conclusions
The ripening time required to achieve complete manganese removal with (reference) full scale filters with
virgin sand and virgin anthracite/sand filter media at two GWTPs; De Punt (The Netherlands) and
Grobbendonk (Belgium) was found to be 55 and 16 days, respectively.
Differences in duration of ripening times between filters of the two GWTPs is caused by a combination of
factors including the different composition of feed water (pH, redox potential, concentration of Fe2+ and
NH4+), applied process design and operational conditions (e.g., iron load, intensity and frequency of
backwashing and physical properties and composition of virgin filter media).
Batch adsorption experiments demonstrated that both (fresh) MOCS and (fresh) MOCA adsorb Mn2+.
Based on Freundlich adsorption isotherm measurements, the manganese adsorption capacity (qe) expressed
per unit weight of adsorbent of MOCA was approximately twice the capacity of MOCS. However, when
expressed per unit volume of adsorbent, which is more relevant for a practical application, the manganese
adsorption capacities of MOCA and MOCS were similar. The adsorption capacity of commercial manganese
adsorbent (Aquamandix), expressed per unit volume, was found to be approximately 3 times higher.
Aging and drying of MOCS, most probably resulted in the loss of manganese adsorption capacity. Besides,
drying of MOCS may have caused the loss of auto-catalytic activity by changes in its structure and complete
manganese oxidation. Finally the biological activity may have been lost.
102
The ripening time of a full scale filter with virgin anthracite/sand filter media, before reaching an effective
manganese removal at GWTP Grobbendonk, of typically 16 days could be eliminated when a 0.10 to 0.30 m
deep layer of fresh MOCA is placed on top of the virgin anthracite/sand filter bed. Because of operational
aspects, it is advisable to apply a MOCA layer with a thickness ≥ 0.2 m.
In a follow up research, based on the results obtained from this study, the effect of water quality and
operational conditions, as well as the role of micro-biology on filter media ripening, will be investigated in
more detail under comparable conditions.
This research is financially and technically supported by WLN and the Dutch water companies Waterbedrijf
Groningen and Waterleiding Maatschappij Drenthe. The authors would like to thank Mr A.A.S. Al Abri
(MSc graduate at UNESCO-IHE) for his contribution to this work. Thanks also to the Belgian water
company Pidpa, for providing a full scale test location Grobbendonk and their willingness to share the data
from their groundwater treatment plant.
6.7 References
Appelo, C.A.J. & Postma, D. 2005. Geochemistry, groundwater and pollution, 2nd edition, CRC press, Boca Raton
(FL), USA.
Bruins, J.H., Petrusevski, B., Slokar, Y.M., Kruithof, J.C. & Kennedy, M.D. 2014a. Manganese removal from
groundwater: characterization of filter media coating. Desalination and Water Treatment, in press.
Bruins, J.H., Vries, D., Petrusevski, B., Slokar, Y.M. & Kennedy, M.D. 2014b. Assessment of manganese
removal from over 100 groundwater treatment plants. Journal of Water Supply: Research and TechnologyAQUA, 63(4), 268-280.
Buamah, R. 2009. Adsorptive removal of manganese, arsenic and iron from groundwater. PhD thesis,
University Wageningen and UNESCO-IHE Delft, The Netherlands, ISBN 978-0-415-57379-5.
Buamah, R., Petrusevski B., de Ridder, D., van de Watering, T.S.C.M. & Schippers, J.C. 2009a. Manganese
removal in groundwater treatment: practice, problems and probable solutions. Water Science and
Technology: Water Supply, 9(1), 89-98.
Buamah, R., Petrusevski, B. & Schippers, J.C. 2008. Adsorptive removal of manganese (II) from the aqueous
phase using iron oxide coated sand. Journal of Water Supply: Research and Technology-AQUA, 57(1), 1-11.
Burger, M.S., Mercer, S.S., Shupe, G.D. & Gagnon, G.A. 2008. Manganese removal during bench-scale
biofiltration. Water Research, 42, 4733-4742.
Cools, B. 2010. Vlaamse Maatschappij voor Watervoorziening (VMW), Belgium. Personal communication.
103
Diem, D. & Stumm W. 1984. Is dissolved Mn2+ being oxidized by O2 in absence of Mn-bacteria or surface
catalysts. Geochimica et Cosmochimica, 48, 1571-1573
Fiol, N. & Villaescusa, I. 2009. Determination of sorbent point zero charge: usefulness in sorption studies.
Environmental Chemistry Letters, 7, 79-84.
Flemming, H.C., Steele, H., Rott, U. & Meyer, C. 2004. Optimirung der in-situ reaktortechnologie zur
dezentralen trinkwassergewinnung und grundwasseraufbereitung durch modelhafte untersuchungen
beteiligter biofilme. Report by the Institute for Sanitary Engineering, Water Quality and Solid Waste
Management of the University of Stuttgart.
Hu, P-Y., Hsieh, Y-H., Chen, J-C. & Chang, C-Y. 2004a. Adsorption of divalent manganese ion on
manganese coated sand. Journal of Water supply: Research and Technology – AQUA, 53(3), 151-158.
Hu, P-Y., Hsieh, Y-H., Chen, J-C. & Chang, C-Y. 2004b. Characteristics of manganese-coated sand using
SEM and EDAX analysis. Journal of Colloid and Interface Science, 272, 308-313.
Katsoyiannis, I.A. & Zouboulis, A.I. 2004. Biological treatment of Mn (II) and Fe (II) containing
groundwater: kinetic considerations and product characterization. Water Research, 38, 1922-1932.
Kim, J. & S. Jung 2008. Soluble manganese removal by porous media filtration. Environmental Technology,
29(12), 1265-1273.
Kim, W.G., Kim, S.J., Lee, S.M. & Tiwari, D. 2009. Removal characteristics of manganese-coated solid
samples for Mn (II). Desalination and Water Treatment, 4, 218-223.
Knocke, W.R., Van Benschoten, J.E., Kearny, M.J., Soborski, A.W. & Reckhow, D.A. 1991. Kinetics of
manganese and iron oxidation by potassium permanganate and chlorine dioxide. Journal of AWWA,
June, 80-87.
Krull, J. 2010. Stadtwerke Emden (SWE), Germany. Personal communication.
NEN 2004 Water quality - Application of inductively coupled plasma mass spectrometry (ICP-MS)
- Part 2: Determination of 62 elements, NEN-EN-ISO 17294-2.
Post, J.E. 1999. Manganese oxide minerals: Crystal structures and economic and environmental significance.
Sahabi, D.M., Takeda, M., Suzuki, I. & Koizumi, J-I. 2009. Removal of Mn2+ from water by “aged” biofilter
media: The role of catalytic oxides layers. Journal of Bioscience and Bioengineering, 107(2), 151-157.
Scherer, E. & Wichmann, K. 2000. Treatment of groundwater containing methane – combination of the
processing stages desorption and filtration. Acta Hydrochemica et Hydrobiologica, 28(3), 145-154.
Stumm, W. & Morgan, J.J. 1996. Aquatic Chemistry, Chemical Equilibria and Rates, 3rd edition, Wiley, New York.
Tebo, B.M., Marger, J.R., Clement, B.G., Dick, G.J., Murray, K.J., Parker, D., Verity, R. & Webb, S.M. 2004.
Biogenic manganese oxides: Properties and mechanisms of formation. Annual Review of Earth and
Planetary Sciences, 32, 287-328.
104
Vandenabeele, J., De Beer, D., Germonpré, R. & Verstreate, W. 1992. Manganese oxidation by microbial
consortia from sand filters. Microbial Ecology, 24, 91-108.
Vandenabeele, J., Van de Woestyne, M., Houwen, F., Germonpré, R., Vandesande, D. &Verstreate, W.
1995. Role of autotrophic nitrifiers in biological manganese removal from groundwater containing
manganese an ammonium. Microbial Ecology, 28, 83-98.
105
Figure: Filter backwashing (key factor controlling start of fast filter media ripening) at GWTP De
Groeve (WBG) - (photo made by K. Borger, WLN, 2015)
106
7
FACTORS CONTROLLING THE RIPENING OF
MANGANESE REMOVAL FILTERS IN COVENTIONAL
AERATION-FILTRATION GROUNDWATER TREATMENT
Main part of this chapter was submitted as:
Kruithof, Maria D. Kennedy (2016). Factors controlling the ripening of manganese removal filters in
conventional aeration-filtration groundwater treatment. Submitted to Desalination & Water Treatment.
107
7.1 Abstract
Relatively long operational time is required to achieve effective manganese removal in conventional
aeration-filtration groundwater treatment with virgin filter media. Ripening period depends on water quality,
operational parameters, and the filter media used. This study assessed the role of filter media type,
backwashing and iron loading on the time required to achieve very effective manganese removal. Filter runs
were conducted with two set-ups each with six pilot filters with virgin sand or anthracite, and different types
of manganese oxide coated sand / anthracite (MOCS/MOCA). Pre-treated groundwater (aeration-rapid
sand filtration), either directly, or after an additional pre-treatment (ultrafiltration-UF), was used as feed
water. UF pre-treatment eliminated head loss development in pilot filters and backwashing was consequently
not required. Filters that received feed water without UF pre-treatment required backwashing after 14 days
of continuous operation. Use of virgin sand and anthracite resulted in comparable ripening time (25 days
and 14 days for feed water without and with UF pre-treatment, respectively). Use of fresh MOCS/MOCA
directly taken from operational filters, eliminated the need for ripening of virgin filter media, while dry
MOCS was less effective than fresh one, while the total period required to achieve highly effective
manganese removal (≥95%) was not shortened.
Keywords: filter backwashing, filter media ripening (manganese oxide coated sand/manganese oxide
coated anthracite), groundwater quality, groundwater treatment, manganese removal.
7.2 Introduction
In many European countries the removal of manganese from groundwater is predominantly achieved by
aeration-rapid sand filtration (Bruins et al., 2014). This treatment approach requires no chemicals (to oxidize
Mn2+), in contrast to manganese removal based on oxidation-filtration that is commonly applied in US and
some other countries (Knocke et al., 1991), is easy to operate, and very cost-effective. Application of this
process, is, however, associated with a number of challenges including the long ripening period of virgin
filter media required to achieve very effective manganese removal (Buamah et al., 2009; Cools, 2010; Krull,
2010). Several parameters are suggested to influence virgin filter media ripening period, including
groundwater quality, type of filter media, and intensity and frequency of filter backwashing (Bruins et al.,
2014; Paassen van, 2010).
Removal of manganese (Mn2+) in aeration-rapid sand filtration, which is often believed to be an autocatalytic adsorption-oxidation process (Graveland and Heertjes, 1975; Stumm and Morgan, 1996), can be
supported by biological manganese oxidation (Vandenabeele et al., 1992; Tebo et al., 2004; Barger et al., 2009;
Burger et al., 2008). Formation of manganese oxide (MnOx) is likely initiated biologically, and over a
prolonged filter run time, MnOx becomes of predominantly physico-chemical origin (Bruins et al., 2015a).
Many researchers have shown that the removal of dissolved manganese (Mn2+), can be facilitated by
adsorption on manganese oxide coated filter media (Hu et al., 2004a,b; Kim and Jung, 2008; Kim et al., 2009).
Filter media suggested to be the most efficient is manganese oxide coated sand / anthracite – MOCS /
MOCA (Olanczuk-Neyman and Bray, 2000; Stembal et al., 2005; Tiwari et al., 2007; Buamah et al., 2008;
Tekerlekopoulou and Vayenas, 2008; Sahabi et al., 2009; Islam et al., 2010; Bruins et al., 2015b). Addition of
a layer of fresh MOCA in the filter with virgin anthracite was reported to completely eliminate filter media
ripening time, while at an another testing location provision of a layer of dried MOCS did not result in any
reduction (Bruins et al., 2015c). It was hypothesized that differences in MOCS / MOCA performance at
108
two testing locations could be explained by dissimilar conditions, under which the tests were carried out,
such as water quality, mode of backwashing and different characteristics of manganese oxide coated media.
The aim of this research was to study the effect of backwashing, and, the type of virgin filter media on the
duration of the ripening period required to achieve very effective (> 90%) manganese removal. In addition,
the potential of freshly taken manganese coated filter media, from very effective operating manganese
removal filters (MOCS as well as MOCA), to reduce ripening period of manganese removal filters with
virgin filter media was investigated. Also dried manganese coated media (with a potential loss of biological
activity and adsorptive properties) were tested.
Manganese oxide coated materials used in this study were obtained from three full scale aeration-filtration
groundwater treatment plants (GWTPs). Figure 7.1 gives an overview of the process schemes of the
GWTPs, and the filters from which Manganese Oxide Coated Sand (MOCS) / Manganese Oxide Coated
Anthracite (MOCA) samples were taken.
Grobbendonk (BE)
Onnen (NL)
De Punt (NL)
pre-aeration
pre-aeration
pre-aeration
rapid sand filtration
(MOCS2)
pH correction/post-aeration
post-aeration
post-aeration
(MOCA)
(MOCS1)
Figure 7.1: Treatment schemes of GWTPs and filters from which MOCS/MOCA samples were
taken.
Table 7.1 gives an overview of quality of feed water to filters, from which MOCS and MOCA were taken.
109
Table 7.1:Quality of feed water to filters, from which MOCS samples were taken
Parameter
Unit
Grobbendonk
Onnen
De Punt
Iron
mg/L
0.09 (0.03 - 0.1)
0.11 (0.02 - 0.5)
5.9 (5.2 - 6.4)
Manganese
mg/L
0.13 (0.10 - 0.15)
0.06 (0.03 - 0.15)
0.23 (0.19 - 0.27)
Ammonium
mg/L
0.08 (< 0.02 - 0.20)
0.06 (0.05 - 0.09)
0.46 (0.34 - 0.61)
pH
[-]
7.6 (7.7 - 7.9)*
7.5 (7.4 - 7.6)
7.1 (7.0 - 7.1)
Oxygen
mg/L
9 (8 - 9.5)
8 (6.8 - 9.1)
10 (9.3 - 10.6)
* pH mainly found to be between 7.5 and 7.6
Fresh MOCA and MOCS1 were taken from well operating filters (manganese removal ≥95%), and directly
(without drying) used in pilot filter columns. In addition, two batches of dried MOCS1 and MOCS2 (taken
out of manganese removal filters and subsequently dried in open air, and stored for 2 and 6 months,
respectively), were also used in this research. Dried coated filter media was included in the research having
in mind that it is not always possible to obtain fresh manganese coated media that can be used for a startup of manganese removal filters with virgin filter media. Prior to using, all filter media samples were rinsed,
to flush out fines.
The pilot filters operated at GWTP Grobbendonk (Belgium). The experimental set-up (Fig. 7.2) consisted
of two sets each with 6-columns installed in a parallel.
Feed water
Ultrafiltration
30
30
80
30
50
50
80
50
sand
reference
(A1)
30
anthracite
reference
(A2)
virgin sand
50
× 4 test
filters
(A3-A6)
sand
reference
(B1)
virgin anthracite
set A
Figure 7.2: Pilot set up at GWTP Grobbendonk
anthracite
reference
(B2)
× 4 test
filters
(B3-B6)
Mn coated material
set B
110
Based on full scale experiments in Grobbendonk, a filter media layer of 30 cm was used. The pilot filter
columns had a diameter of 10 cm, and the filter bed length was 80 cm, composed of 50 cm support material
(virgin sand) and 30 cm virgin sand, anthracite, MOCA or MOCS as follows:
•
•
•
•
•
•
A1 / B1: virgin sand (reference for MOCS layers);
A2 / B2: virgin anthracite (reference for MOCA layers);
A3 / B3: MOCA (fresh);
A4 / B4: MOCS1 (fresh);
A5 / B5: MOCS1-dried;
A6 / B6: MOCS2-dried.
The physical properties of MOCS and virgin sand/anthracite, the chemical composition of the coating and
the Freundlich adsorption isotherm constants for Mn2+ adsorption on MOCS1, MOCS2 and MOCA (all
fresh) are given in Table 7.2.
Supernatant layer of 0.3 m was provided above filter media. The pilot filter columns were operated in down
flow mode at constant filtration rate of 5.1 ± 0.5 m/h. The columns in test set A were backwashed after
approximately every 2 weeks of continuous operation. Backwashing was carried out with water only, at a
backwash rate of 30-35 m/h, resulting in approximately 10-20 % filter bed expansion, with typical duration
of 10 min. The columns in the test set B, that received water after additional UF pre-treatment, were not
backwashed during the whole study period, because no increase of head loss was observed
Table 7.2: Physical properties of fresh MOCS, MOCA, and virgin sand /anthracite, chemical
composition of the coating and the Freundlich adsorption isotherm constants for Mn 2 + adsorption on
fresh MOCS and MOCA
Parameter
Unit
MOCS1
MOCS2
MOCA
Virgin
Virgin
sand
anthracite
Grain size (d10-d90)
mm
1.3 - 2.0 1.6 - 3.1 0.8 - 1.5 0.4 - 0.8
0.8 - 1.6
Uniformity
[-]
1.21
1.58
1.52
1.30
1.42
coefficient
Bulk density
Kg/L
1.376
1.177
0.650
1.459
0.635
Particle density
Kg/L
2.332
2.326
1.176
2.586
1.400
Porosity
%
46.0
49.4
44.7
43.6
54.6
pHPZC
[-]
7.8
7.2
8.0
6.1 - 6.5
9.1
Coating composition
Mn
mg/g
30.4
12.8
13.5
Fe
mg/g
21.4
158
2.2
Ca
mg/g
7.8
8.9
2.5
Al
mg/g
2.4
0.5
0.3
Freundlich adsorption
isotherm constants
(mg/g)/(mg/L)
K
0.70
0.45
0.91
1/n
[-]
1.25
1.31
1.34
r2
[-]
0.97
0.91
0.91
qe at Ce = 0.2 mg/L
mg/g
0.193
0.132
0.276
qe at Ce = 0.2 mg/L
g/L
0.265
0.155
0.179
The filtrate from the first stage of the full scale GWTP Grobbendonk was used directly, or after
ultrafiltration -UF (Inge, dizzer 500SB, pore size 0.02 µm), as feed for the pilot filter columns. UF filtration
111
was applied to retain particles present in the feed water. It was assumed that ultrafiltration would prevent,
or at least strongly reduce, head loss development in the filter bed, with associated reduced backwashing
frequency. Filtrate of the first filtration step is also feed to full scale manganese removal filters at this plant
(Table 7.1). Composition of pilot feed water is given in Table 7.3.
Manganese in feed water was in dissolved (Mn2+) form, while iron was present predominantly as (hydr)oxide.
Samples of feed water and filtrate (a sampling point just below the top layer of 30 cm), were taken daily, and
the manganese concentration was analysed with ICP-MS.
Table 7.3: Quality (average and range) of feed water to pilot filters, with and without UF pretreatment
Parameter
Unit
Without UF
With UF
Iron
mg/L
0.09 (0.03 - 0.1)
< 0.02
DOC
mg/L
1.58 (1.51 - 1.68)
1.46 (1.33 - 1.60)
Manganese
mg/L
0.13 (0.10 - 0.15)
Ammonium
mg/L
0.08 (< 0.02 - 0.20)
pH
[-]
7.6 (7.5 - 7.9)*)
Oxygen
mg/L
9 (8 - 9.5)
Redox potential
mV
+250 (+200 - +290)
* ) most of the time pH ranged from 7.5 to 7.6
7.4.1 Filter ripening with virgin media and the effect of filter
backwashing
100
80
60
40
20
0
0
5
10
15
20
virgin sand (A1)
virgin anthracite (A2)
backwash
25
Fig. 7.3 shows the comparison of virgin sand and virgin anthracite ripening time, and the effect of UF pretreatment on the ripening period during the first 25 days of filter run with virgin filter media.
100
80
60
40
20
0
0
5
10
15
20
25
virgin sand (B1)
virgin anthracite (B2)
Figure 7.3: Mn removal efficiency (%) as a function of the filter ripening time, type of virgin filter
media and UF pre-treatment; feed water without (left), and with UF pre-treatment (right)
From Fig. 7.3 it can be seen that ripening time required to achieve effective manganese removal with virgin
sand and anthracite, was similar when the same feed water pre-treatment was applied. The results imply that
112
different physical properties of sand and anthracite (e.g. particle size and pHPZC, see Table 7.2.) did not
significantly affect Mn2+ adsorption, and related rate of Mn-coating development. However, UF pretreatment of feed water (Fig. 7.3 right) resulted in faster filter media ripening. Highly effective manganese
removal (≥95%) was achieved after approximately 14 days, as compared to 24 days in the filters when no
UF was applied. The difference in media ripening time of the two set-ups could be presumably attributed
to the effect of backwashing on development of the Mn-coating (Bruins et al., 2015a) . Mn-coating is known
to be essential for effective manganese removal in aeration-rapid sand filtration treatment (Hu et al., 2004a,b;
Kim and Jung, 2008; Kim et al., 2009). In general, it is known that the oxidation-reduction potential (ORP)
influences the manganese removal efficiency. However, in this research the influence of ORP was limited.
During the whole test, ORP was close to around + 250 mV. Also after each backwashing.
During the first two weeks no filter backwash was applied (head loss developed in filter bed was
compensated by opening and adjusting a flow control valve to allow operation of pilot column at constant
filtration rate) and the ripening of all filters with virgin filter media in both set-ups was very similar,
irrespective of applied pre-treatment. Backwashing of filters A1 and A2 conducted after 14 days of
continuous operation, however, resulted in reduction of the Mn2+ removal efficiency from 28 % to <10 %
(Fig 7.3 left). Due to the UF pre-treatment, the feed water of columns B1 and B2 contained no particulate
matter and as a consequence no filter backwashing of columns B1 and B2 was required (Fig 7.3 right).
Although the backwashing used in this study was gentle and performed only after 14 days (Fig. 7.3 left, fine
dotted line), the observed differences in Mn removal efficiency suggest that backwashing has a substantial
effect on the ripening time of filter media.
Negative effect of backwashing during the ripening phase on manganese removal could be likely attributed
to:
• (partial) removal of bacteria responsible for oxidation of Mn2+ that developed on virgin filter media
during ripening period (Vandenabeele et al., 1992; Tebo et al., 2004; Barger et al., 2009; Burger et al.,
2008; Bruins et al., 2015a) and / or
• (partial) removal of freshly formed MnOX (Birnessite), which has highly autocatalytic properties to
adsorb and subsequently oxidize Mn2+ (Bruins et al., 2015b; Post, 1999).
Fig. 7.4 shows the effect of the filter backwashing on manganese removal during the total testing period of
approximately 70 days. Manganese concentration in filtrate was analysed before and after each backwash
cycle.
113
100
80
60
40
20
0
0
10
20
30
40
50
virgin sand
60
70
backwash
Figure 7.4: Effect of filter backwashing on manganese removal efficiency during filter media ripening
with virgin sand.
From Fig. 7.4 it can be seen that directly after each backwash cycle, the manganese removal efficiency
decreased. The decrease was most pronounced after the first backwash cycle (manganese removal reduced
from 28 % to <10 %), carried out 14 days after the start of the filter operation, when the manganese removal
was still rather ineffective presumably due to only limited partial media coating with birnessite (Bruins et al.,
2015b). Backwashing at that stage of ripening period likely partially removed initial birnessite deposits, and
bacteria responsible for Mn2+ oxidation, and, consequently had a strong negative impact on the Mn removal
efficiency. After the second and the third backwash cycles (performed after 28 and 42 days of filter
operation), the reduction of manganese removal efficiency was less pronounced, but still obvious - from
94 % to 84 % after the second backwash cycle, and from 98 % to 93 % after the third backwash cycle. After
the fourth backwash cycle (57 days of the filter operation), no significant decrease of manganese removal
efficiency was observed, showing that there was sufficient birnessite on filter media, and partial removal due
to backwashing did not significantly hampered manganese removal.
Filters treating iron and manganese containing groundwater need to be periodically backwashed to remove
particles caught in filter bed voids, causing head loss development. The majority of these particles are iron
(hydr) oxides formed by oxidation of dissolved iron with oxygen. The backwashing of pilot filters in this
study was carried out with a very low frequency (approximately once in 14 days) due to low iron
concentration in the feed water (up to a maximum of 0.1 mg/L for the set-up A), and as a consequence,
iron loading in the columns was limited (approximately 0.1 kg Fe/m2 filter area, per filter run). In practice,
however, iron loading can be much higher, introducing the need for more frequent backwash cycles (e.g.,
the backwash frequency of GWTP Noordbargeres, water company Drenthe (NL), is more than once per
day, because of high iron concentration in the feed water of approximately 14 mg/L, with iron loading of
1.35 kg Fe/m2/filter run). Under such conditions, the negative effect of backwashing on duration of
ripening period required to achieve very effective manganese removal with virgin filter media will be much
more pronounced (very effective manganese removal in the filter is typically achieved after more than four
months).
In Fig. 7.5 the filter media ripening time of the pilot filter from this study was compared with the filter
ripening time of the full scale filter GWTP Grobbendonk Bruins et al., 2015c). In both cases the virgin
anthracite was used as filter media. Feed water quality for both filters was the same (Table 7.3, column
"without UF"). However, the applied filtration rate, and consequently iron loading and backwash frequency
114
were, different. In addition, the depth of the anthracite layer was different (0.5 m and 0.3 m in the full scale
and the pilot filters, respectively). Inspection of the anthracite layer after completion of filter run showed
that the manganese was removed mainly in the top 0.10-0.15 m of the filter, suggesting that different heights
of the anthracite layer did not affect the results.
100
80
60
40
20
0
0
10
20
30
40
full scale
pilot
Figure 7.5: Comparison of the Mn removal efficiency as a function of the filter ripening time for full
scale and pilot filters with virgin anthracite (filtration rate: Vf = 5.1 m/h pilot filter and Vf = 2.5
m/h full scale filter).
Highly effective manganese removal (≥ 95%) was achieved after approximately 16 days in the full scale
plant, and after 24 days in the pilot filter column. The difference in applied filtration rate (2.5 m/h and
5.1 m/h, for full scale and pilot filters, respectively), resulted in different iron loadings
(0.05 and 0.1 kg Fe/m2 filter area, per filter run). As a consequence, the pilot filter column had to be
backwashed earlier, than the full scale filter (14 and 30 days, respectively). The first backwashing of the pilot
column took place just when manganese removal efficiency started to rapidly increase (black arrow in
Fig. 7.5), while for the full scale filter, the first backwashing was applied only after 30 days, when the highly
effective manganese removal was already achieved.
Iron concentrations in feed water of drinking water treatments plants may range from < 0.02 mg/L to more
than 30 mg/L, with related iron loading from less than 0.01 to over 10 kg/m2 filter area, per filter run
(Bruins et al., 2014). As a consequence, the backwash frequency in practice can vary between once per month
to more than once per day. Findings emerging from this study, suggest that the ripening time required to
achieve very effective manganese removal with virgin filter media is strongly affected by the applied
backwash frequency that is correlated to the iron loading. High iron loading, and consequently high
backwash frequency, could be reduced by operating filters with lower filtration rate during the ripening
period. Another option is to recirculate part of the filtrate, thus lowering the iron loading and consequently
the backwash frequency. The reduction of the plant operational capacity during the ripening period will in
most cases not be a problem because water with high manganese concentration, has to be disposed anyway.
Backwashing with water only is recommended during the ripening period to limit removal of bacteria and
MnOx deposits developed on the filter media, since backwashing with air and water, which is normally used,
is much more abrasive.
115
7.4.2 Filter ripening with addition of a layer of manganese coated
media
100
80
60
40
20
0
0
5
10
15
20
MOCA (A3)
MOCS1 (A4)
backwash
25
Fig. 7.6 shows the manganese removal efficiency of pilot filters with virgin sand containing a layer of fresh
MOCA or MOCS1 during the initial 2-3 weeks of operation.
100
80
60
40
20
0
0
5
10
15
20
MOCA (B3)
MOCS1 (B4)
25
Figure 7.6: Mn removal efficiency (%) as a function of the filter run time for pilot filters containing a
layer of fresh MOCS 1 or MOCA; feed water without (left) and with UF (right) pre-treatment.
Fig. 7.6 shows that the addition of a layer of either fresh MOCA or MOCS1 on top of the virgin sand
resulted in very effective (>90 %) manganese removal already after one day of filter operation, irrespective
of UF pre-treatment. It has been shown, that birnessite presence in filter media coating is essential for an
effective removal of dissolved manganese (Bruins et al., 2015b). The results from this study show that MOCS
or MOCA, freshly taken from well performing manganese removal filters were auto-catalytically active. This
assumption was supported by calculation of the maximum adsorption capacity of the MOCS/MOCA layers
(2.35 litres). Based on the adsorption isotherms (Table 7.2) Mn2+ adsorption capacity (at Ce=0.2 mg/L) was
calculated to be 623, 365 and 420 mg Mn2+ for MOCS1, MOCS2 and MOCA, respectively. Given the daily
filtrated volume through each column (960L) and the average Mn2+ concentration in the feed water
(0.13mg/L) the adsorption capacity of MOCS layers expected to be exhausted after 5.0; 2.9 and 3.4 days,
for MOCS1, MOCS2 and MOCA, respectively. Fig. 6 shows that manganese removal was still very effective
after 25 days of continuous pilot filters operation. This strongly suggests the presence of birnessite, that
enables an efficient auto-catalytic adsorption and subsequent oxidation of Mn2+.(Bruins et al., 2015b; Post,
1999). Therefore replacing the top layer of a new (virgin) filter by ‘active’ manganese coated media could
eliminate the ripening time of manganese removal filters.
Fig. 7.6 also shows that the manganese removal efficiency of fresh MOCA and MOCS1 was almost identical,
as expected based on the Freundlich adsorption isotherm constants (Table 7.2). Use of fresh MOCA or
MOCS, from well performing manganese filters (even with different physical properties and chemical
composition) can consequently reduce, or even eliminate the long filter ripening period.
Results obtained (Fig 7.6 left) also confirmed that backwashing has rather limited effect on manganese
removal for filters media with well developed (birnessite) coating.
Fig. 7.7 shows the manganese removal during initial 2-3 weeks of operation of pilot filter columns with
virgin sand and a layer of dry MOCS1 and MOCS2, or a layer of fresh MOCS1. Feed water without (Fig. 7.7,
left) and with UF pre-treatment (Fig. 7.7, right) was used in these filter runs.
116
100
80
60
40
20
0
0
5
10
15
20
MOCS1 (A4)
MOCS1-dry (A5)
MOCS2-dry (A6)
backwash
25
100
80
60
40
20
0
0
5
10
15
20
MOCS1 (B4)
MOCS1-dry (B5)
MOCS2-dry (B6)
25
Figure 7.7: The Mn removal efficiency as a function of filter run time for pilot filters containing
layers of dried or fresh MOCS; feed water without (left) and with UF (right) pre-treatment
Fig. 7.7 shows that the performance of the filters containing a layer of dry manganese coated media was not
as efficient as the filters with a layer of a fresh manganese coated media. The difference was more
pronounced when no UF pre-treatment was applied (Fig 7.7 left). Compared to the fresh material, the
removal of manganese was less efficient from the start of the filter run of the pilot filters with a layer of
MOCS2-dry, while for pilot filters with MOCS1-dry showed only during the initial 5 days of operation very
effective Mn removal (>95%), that was subsequently reduced to approximately 80% and 90% for feed water
without and with UF pre-treatment, respectively. For MOCS2-dry the Mn removal efficiency started strongly
decreasing from the 3rd day of filter run.
Dry MOCS is less effective for manganese removal due to several reasons. Firstly, MOCS during drying
process may have lost part of its adsorptive capacity, presumably due to oxidation of autocatalytic active
MnOX (birnessite) into less autocatalytic active MnOX (pyrolusite). As a result, the number of available
adsorptive sites decreases (Post, 1999). Secondly, drying the MOCS can have negative effect on the structure
of the manganese oxide. The auto-catalytically active birnessite, consisting of plates (Post, 1999), may
irreversibly collapse during drying, subsequently additionally limiting the number of available adsorption
sites. Finally, filter media drying might cause a loss of the biological activity on the media surface due to
manganese oxidizing bacteria die-off. These bacteria likely play an important role in the process of
manganese oxidation and removal (Vandenabeele et al., 1992; Tebo et al., 2004; Barger et al., 2009; Burger et
al., 2008; Bruins et al., 2015a).
Fig. 7.7 also shows that dry MOCS2 removed manganese less effectively than dry MOCS1-dry. The
differences between the two MOCS media could be likely attributed to difference in manganese content
and associated available adsorptive sites (e.g. MOCS2-dry that was less effective for manganese removal
contained 12.8 mg Mn/g, while MOCS1-dry contained 30.4 mg Mn/g). Another explanation for the
differences in removal capacities of the two dry MOCS media might be the storage time of two materials;
MOCS1-dry was stored for 2 months, whereas MOCS2-dry was stored for 6 months before their use in this
study. Longer storage (in a dry air) presumably resulted in more pronounced loss of auto-catalytic properties
through mechanisms earlier discussed.
Backwashing of the pilot filter with a layer of dry MOCS1 and MOCS2 had a short positive effect on
manganese removal that could be likely explained by removal of iron precipitates that physically blocked
adsorption sites on filter media.
117
In both pilot filters with dry manganese coated media, irrespective of the UF pre-treatment, manganese
removal efficiency started to improve after 12 days of pilot filter operation. At approximately the same time
manganese removal started increasing in pilot filters with virgin sand or anthracite (Fig. 7.3). The coinciding
ripening times indicate that dry manganese coated media behaved similar to virgin filter media, likely due to
formation of manganese oxides and a biologically active layer that were required to facilitate effective
manganese removal.
The findings of this research can be used in practice: addition of a layer (≥0.3m) of fresh manganese coated
media, strongly reduces or completely eliminates the filter ripening time. Manganese coated media can easily
be taken from a well-operating manganese removal filter and transferred to a filter with virgin media. This
procedure has to be done only once, and there is no need for follow up replacements of manganese coated
filter media.
Filters are immediately ready for operating after using a fresh layer of MOCS or MOCA. As a consequence
of this fast filter media ripening procedure, a water company saves operational costs (saving water, labor,
costs for analysis, no need for extra filters, etc.). Savings are strongly depending on the original ripening
time, without the use of manganese coated filter media.
If no fresh layer of manganese oxide coated filter media from a well-operating manganese removal filter can
be used in practice, the long ripening period of manganese removal filters with virgin filter media can be
reduced by operating filters, temporarily, at a lower filtration rate during the ripening period, consequently
reducing the iron loading and thus limiting the backwashing frequency. Although in this situation the water
production capacity of the filter is, temporarily, also reduced, possible advantages are more pronounced.
7.5 Conclusions
Findings emerging from this study focused on the investigation of key factors controlling the ripening period
of manganese removal filters with virgin media in conventional aeration-filtration treatment lead to the
following conclusions:
•
•
•
•
•
•
Despite different physical properties, virgin sand and virgin anthracite will have similar ripening
time required to achieve very effective manganese removal, assuming that the feed water quality
and operational conditions are identical.
Filter backwashing prolongs the ripening time of a virgin filter. The influence of backwashing
becomes less pronounced with the progress of filter media coating, due to development of thicker
layer of biomass and / or auto-catalytically active birnessite on adsorption media surface.
Backwashing has no, or only limited influence, on manganese removal when a layer of fresh
manganese oxide coated filter media from well performing manganese removal filters, is added to
a filter bed with virgin media.
Ripening of manganese removal filters with virgin filter media can be shortened by temporarily
operating filters at lower filtration rate with associated less frequent backwashing cycles.
Addition of a layer (≥0.3m) of fresh manganese coated media, from a well-operating manganese
removal filter, to a filter with virgin media, strongly reduces or completely eliminates the filter
ripening time.
Compared to fresh-, dry MOCS has inferior manganese removal properties due to lower adsorption
capacity and likely very limited potential for catalytic Mn2+ adsorption-oxidation.
118
•
•
Time required to create a new (bio-active) auto-catalytically layer of birnessite on the surface of dry
manganese coated or virgin filter media is similar.
The long ripening period of manganese removal filters with virgin filter media in practice, can be
reduced or even eliminated by (1) addition of a layer of fresh MOCA or MOCS from a wellfunctioning manganese removal filters, and (2) operating filters at lower filtration rate during the
ripening period consequently reducing the iron loading and the backwashing frequency.
This research was financially and technically supported by WLN and the Dutch water companies Groningen
(WBG) en Drenthe (WMD). The authors wish to thank water company Pidpa (Belgium), for their
willingness to share the data from their groundwater treatment plant and for the opportunity to carry out
pilot tests at their Grobbendonk facilities. Special thanks are extended to Mrs. Ann Maeyninckx and Mrs.
Martine Cuypers (Pidpa) for helping run the pilot and performing countless analyses.
7.7 References
Barger J.R., Fuller C.C., Marcu M.A., Brearly A., Perez De la Rosa M., Webb S.M., Caldwell W. A. 2009.
Structural characterization of terrestrial microbial Mn oxides from Pinal Ckeek. AZ. Ceochimica et
Cosmochimica Acta, 73, 889-910.
Bruins J. H., Vries D., Petrusevski B., Slokar Y.M., Kennedy M.D. 2014. Assessment of manganese removal
from over 100 groundwater treatment plants. Journal of Water Supply: Research and Technology –
AQUA, 63 (4) 268-280.
Bruins J.H., Petrusevski B., Slokar Y.M., Huysman K., Joris K., Kruithof J.C., Kennedy M.D. 2015a.
Biological and physico-chemical formation of Birnessite during the ripening of manganese removal
filter, Water Res. 69, 154-161.
Bruins J.H., Petrusevski B., Slokar Y.M., Kruithof J.C., Kennedy M.D. 2015b. Manganese removal from
groundwater: characterization of filter media coating. Desalination & Water Treatment, 55 (7), 1851
- 1863.
Bruins J.H., Petrusevski B., Slokar Y.M., Huysman K., Joris K., Kruithof J.C., Kennedy M.D. 2015c.
Reduction of ripening time of full scale manganese removal filters with manganese oxide coated filter
media. Journal of Water Supply: Research and Technology – AQUA, 64 (4), 434-441.
Buamah R., Petrusevski B., Schippers J.C. 2008. Adsorptive removal of manganese (II) from the aqueous
phase using iron oxide coated sand. Journal of Water supply: Research and Technology – AQUA, 57
(1), 1-11.
119
Buamah R., Petrusevski B., de Ridder D., van de Watering S., Schippers J.C. 2009. Manganese removal in
groundwater treatment: practice, problems and probable solutions. Journal of Water Science and
Technology: Water Supply, 9 (1) (2009) 89-98.
Burger M.S., Krentz C.A., Mercer S.S., Gagnon G.A. 2008. Manganese removal and occurrence of
mangnanese oxidizing bacteria in full-scale biofilters, J. Water Supply Res. Technol. AQUA 57.5. 351
- 359.
Cools B. 2010. De Watergroep (Flemish water company), Personal communication, Belgium.
Graveland A., Heertjes P.M. 1975. Removal of mangnanese from groundwater by heterogeneous
autocatalytic oxidation. Trans. Chem. Eng. 53, 154-164.
Hu P-Y., Hsieh Y-H., Chen J-C., Chang C-Y 2004a. Adsorption of divalent manganese ion on manganese
coated sand. Journal of Water Supply: Research and Technology – AQUA, 53 (3), 151-158
Hu P-Y., Hsieh Y-H., Chen J-C., Chang C-Y. 2004b. Characteristics of manganese-coated sand using SEM
and EDAX analysis. Journal of Colloid and interface science, 272, 308-313.
Islam A.A., Goodwill J.E., Bouchard R., Tobiasen J.E., Knocke W.R.. 2010. Characterization of filter media
MnO2(s) surfaces and Mn removal capability. Journal AWWA, 102 (9), 71-83.
Kim J., Jung S. 2008. Soluble manganese removal by porous media filtration. Environmental Technology,
29 (12), 1265-1273.
Kim W.G., Kim S.J., Lee S.M., Tiwari D. 2009. Removal characteristics of manganese-coated solid samples
for Mn(II). Desalination and Water Treatment, 4, 218-223.
Knocke W.R., van Benschoten J.E., Kearny M.J., Soborski A.W., Reckhow D.A. 1991. Kinetics of
manganese and iron oxidation by potassium permanganate and chlorine dioxide. Journal of AWWA
June, 80-87.
Krull J. 2010. Stadwerke Emden – SWE (German water company), Personal communication, Germany.
Olanczuk-Neyman K., Bray R. 2000. The role of physico-chemical and biological processes in manganese
and ammonia nitrogen removal from groundwater. J. Pol Environ. Stud., 9 (2), 91-96.
Paassen van J. 2010. Vitens (Dutch water company), Personal communication, The Netherlands.
Post J.E. 1999. Manganese oxide minerals: Crystal structures and economic and environmental significance.
Proceedings of the National Academy of Sciences USA, Vol 96, 3447-3454.
Sahabi D.M., Takeda M., Suzuki I., Koizumi J-I. 2009. Removal of Mn2+ from water by ‘’aged’’ biofilter
media: The role of catalytic oxides layers. Journal of Bioscience and Bioengineering, 107 (2), 151-157.
Stembal T., Markic M., Ribicic N., Briski F., Sipos L. 2005. Removal of ammonia, iron and manganese from
ground waters of Northern Croatia – pilot plant studies. Process Biochemistry, 40, 327-335.
Stumm, W., Morgan J.J. 1996. Aquatic chemistry, Chemical Equilibria and Rates, third ed. Wiley, New York.
120
Tebo B.M., Marger J.R., Clement B.G., Dick G.J., Murray K.J., Parker D., Verity R., Webb S.M 2004.
32, 287-328.
Tekerlekopoulou A.G., Vayenas D.V. 2008. Simultaneous biological removal of ammonia, Iron and
manganese from potable water using a trickling filter. Biochemical Engineering Journal, 39, 215-220.
Tiwari D., Yu M.R., Kim M.N., Lee S.M., Kwon O.H., Choi K.M., Lim G.J., Yang J.K. 2007. Potential
Vandenabeele J., de Beer D., Germonpré R., Verstreate W. 1992. Manganese oxidation by Microbial
consortia from sand filters. Microbial Ecology, 24 (1992) 91-108.
121
Figure: Pilot set up at GWTP De Grobbendonk (water company Pidpa) - (photo made by J. H.
Bruins, WLN, 2012)
122
8
GENERAL CONLUSIONS
123
8.1 Overall conclusions and perspective
The main challenge, regarding manganese removal from groundwater with aeration-rapid sand filtration, is
to get a better understanding of the mechanisms involved in the ripening of virgin filter media and to use
this knowledge to shorten or preferably completely eliminate the long ripening period of virgin media and
to prolong the lifetime of the filter media.
Based on the research described in this thesis it can be concluded that several water quality and operational
parameters, such as iron (and iron loading), pH, ammonia and empty bed contact time are of importance
for efficient manganese removal. Furthermore, Birnessite was identified as the oxide predominantly involved
in effective manganese removal. Birnessite is highly reactive, with outstanding auto-catalytic adsorptive and
oxidative properties. This research clearly demonstrates (at pilot scale) that Birnessite was produced
biologically at the start of filter ripening but as filter ripening progressed, the production of Birnessite was
physico-chemical. During this study, molecular DNA analysis showed a bacteria population shift from the iron
oxidizing genus Gallionella, found in iron removal filters, to the manganese and nitrite oxidizing genus
Pseudomonas and Nitrospira, present in the manganese removal filter. Furthermore, bacteria species that could
potentially oxidise manganese were also identified.
During pilot research, key factors, for efficient start-up of manganese removal filters were defined. Most
important is the influence of backwash frequency on manganese removal efficiency at the start of filter
media ripening. Furthermore it was found that the use of freshly prepared manganese oxide coated media
practically eliminated the ripening period. Optimal conditions to shorten filter ripening (in practice) using
MOCS and MOCA, containing Birnessite, were established.
The findings and conclusions of this research defined important water quality and operational parameters
and knowledge about the role bacteria play in rapidly ripening filter media. Knowing the bacteria involved
and being aware of the damaging effect of filter backwashing, especially at the start of filter ripening, yields
a solid base for a better understanding of the mechanisms involved in speeding up the ripening process with
virgin filter media. This information provides water companies the opportunity to develop an innovative
manganese removal process i.e., ‘a manganese oxidizing bacteria-friendly process’, with a restricted backwash
frequency at the start-up of filter ripening. And if possible using fresh MOCS or MOCA to shorten or
preferably completely eliminate the long ripening period of new filters and prolong the useful lifetime of
the media.
In this chapter each individual aim is discussed in more detail in a separate section to place the objectives in
a broader perspective. Finally a general outlook with some recommendations for use in practice and further
research are given.
124
8.2 Assessment of manganese removal from over 100
groundwater treatment plants
The aim of this study was to make an inventory of water quality and operational parameters affecting
manganese removal through aeration-rapid sand filtration and to establish correlations between these
parameters and manganese removal efficiency.
The focus of the overview was on the manganese removal efficiency in the first filtration stage. Data from
over 100 selected full-scale groundwater treatment plants were collected, and univariate and multivariate
statistical (PCA) analyses conducted. Furthermore, close inspection of the collected data and conducted
statistical analyses (univariate correlations and PCA) showed that the manganese removal efficiency was
influenced by several parameters simultaneously, including both water quality and operational (design)
parameters.
The multivariate statistical method (PCA) revealed that iron loading, NH4+ removal efficiency and pH of
the filtrate played a major role in manganese removal, while oxygen concentration in the filtrate, and
phosphate, manganese and NH4+ concentration in the raw water influenced manganese removal to a lower
degree. Operational parameters, such as filtration rate and empty bed contact time (EBCT) were found to
be of secondary importance.
PCA, Univariate statistics and assessment of available data indicated that a very effective manganese removal
efficiency in the first aeration-filtration stage with simultaneous removal of iron and ammonia, was achieved
under the following conditions:
•
•
•
•
•
•
pH of filtrate
filtration rate
oxygen in filtrate
:
:
:
:
:
:
> 85%
< 2.7 kg Fe/m2
> 7.1
< 10.5 m/h
> 11.5 min
≥ 1 mg/L
The results of this study give directions to engineers to design a traditional aeration-filtration process for
efficient manganese removal. Based on the raw water quality to be treated, and the results of this study the
most suitable filter configuration and the required filter dimensions and filtration conditions can be selected.
In addition, results emerging from univariate statistics and assessment of available data from full scale plants
can help in making a decision if manganese removal can be combined with removal of iron and ammonium
in a single filtration step. Therefore, the amounts of iron, ammonia and manganese to be treated must be
known and based on the conditions at which they can be removed together in one filtration step, provided
by this study, it can be decided if application of one filtration step only is feasible for a new treatment design.
125
8.3 Manganese removal from groundwater: Characterization
of filter media coating
Removal of manganese in conventional aeration-filtration groundwater treatment plants (GWTPs) results
in the formation of a manganese oxide coating on filter media. The formation of this coating is an essential
prerequisite for efficient manganese removal. There are several types of manganese oxides, which have
different affinities for autocatalytic adsorption/oxidation of dissolved manganese. The aim of this study was
to characterize the manganese oxide(s) present on filter media from successfully operating manganese
removal plants. Characterization of filter media samples from full-scale groundwater treatment plants and
identification of manganese species was carried out by X-ray diffraction (XRD), Scanning Electron
Microscopy coupled with Energy Dispersive X-radiation (SEM-EDX), Raman spectroscopy and Electron
Paramagnetic Resonance (EPR).
The Raman spectroscopy, XRD and SEM analyses showed that the manganese oxide in the coating of the
manganese removing filter media is poorly crystalline. Raman spectroscopy and EPR analysis showed that
the predominant manganese oxide, responsible for an effective removal of dissolved manganese, is of a
Birnessite type. Calculation of ΔH and the g factor from EPR analysis and comparison of these parameters
with results from literature identified the Birnessite to be of physicochemical origin, but the sampling after
a ripening period of about 15 years does not exclude that Birnessite formation starts via a biological pathway.
It is generally accepted that the manganese oxidation pathway is via Hausmannite and Manganite. However,
the results transpiring from this research show that in water treatment practice, oxidation of manganese on
the surface of manganese removal filter media causes formation of a Birnessite type of manganese oxide.
Birnessite has very good properties for adsorption and autocatalytic oxidation of dissolved manganese.
The knowledge that Birnessite is the predominant manganese oxide in filter media effectively removing
manganese, may enable shortening of the ripening time in conventional aeration-filtration groundwater
treatment plants by creating conditions that favor the formation of this compound. Another possibility is
using freshly coated filter media containing Birnessitte to enhance the start of the filter media ripening
process. Therefore it is necessary to install the freshly produced Birnessite containing filter media at a filter
height where manganese removal takes place.
8.4 Biological and physico-chemical formation of Birnessite
during ripening of manganese removal filters
The efficiency of manganese removal in conventional groundwater treatment consisting of aeration
followed by rapid sand filtration, strongly depends on the ability of filter media to promote adsorption of
dissolved manganese and its subsequent autocatalytic oxidation. Parallel studies, described in this thesis,
have shown that the compound responsible for the autocatalytic activity in ripened filters is a manganese
oxide called Birnessite. The aim of this study was to determine if the ripening of manganese removal filters
and the formation of Birnessite on virgin sand is initiated biologically or physico-chemically. The ripening
of virgin filter media in a pilot filter column fed by pre-treated manganese containing groundwater was
studied for approximately 600 days. Samples of filter media were taken at regular time intervals, and the
manganese oxides formed in the coating were analysed by Raman spectroscopy, Electron Paramagnetic
Resonance (EPR) and Scanning Electron Microscopy (SEM).
126
These analyses confirmed that during the whole period of filter media ripening, a Birnessite type of
manganese oxide was the predominant mineral in the coating. Furthermore, Raman spectroscopy results
showed that Birnessite was already present in the coating at a very early stage of the ripening process. EPR
analyses and comparison with literature showed that the Birnessite type of manganese oxide, at the
beginning of the ripening process was of biological origin. With the progress of filter ripening and
development of the coating, Birnessite formation became predominantly of physico-chemical origin,
although biological manganese oxidation continued to contribute to the overall manganese removal.
Especially, solids collected from filter backwash water throughout the whole ripening period were
consistently of biological origin, suggesting that biological oxidation of adsorbed manganese was present
throughout the filter run. Finally SEM micrographs, showed a clear difference between biologically and
physico-chemically formed Birnessite. Biologically produced Birnessite is fluffy, plate structured, whereas
physico-chemically produced Birnessite shows a sponge or coral structure.
Figure 8.1: Sponge or coral structure of physico-chemically produced Birnessite (magnification 10,000
x)- Photo made by Arie Zwijnenburg (Wetsus).
The knowledge that manganese removal in conventional groundwater treatment is initiated biologically, may
help reducing long ripening times by creating conditions that are favourable for the growth of manganese
oxidizing bacteria, e.g., by limiting back wash frequency and / or intensity. In practice back wash frequency
can be limited, by reducing the iron loading on the filter. Possibilities to reduce iron loading in practice, are
lowering the filtration rate, partially circulating of filtrate and or feeding the filters with water containing low
amounts of Fe2+ (if available).
127
8.5 Identification of the bacterial population in manganese
removal filters
Fast filter media ripening, for manganese removal in conventional aeration-rapid sand filtration groundwater
treatment, greatly depends on the way autocatalytic adsorptive MnOx is formed. As said before, the oxide
responsible for fast autocatalytic action in ripened filters is (biological formed) Birnessite. The aim of the
study described in this chapter was to identify the bacteria present in recently ripened manganese removal
filters. For this purpose molecular DNA analysis, such as “next generation DNA sequencing”, qPCR and
MALDI-TOF analysis were used.
Based on the results of this study, especially from “next generation DNA sequencing” analyses, a bacteria
population shift was established, from the iron oxidizing genus Gallionella, found in the iron removal filter
to the manganese and nitrite oxidizing genus Pseudomonas and Nitrospira, respectively, present in the
manganese removal filter. However, 47.6% of the bacteria population in the manganese oxidizing column,
belongs to smaller populations or could not be identified. Applying qPCR it was shown that the most
abundant manganese oxidizing genus was Pseudomonas sp. Furthermore it was established that the presence
of the well-known Mn2+-oxidizing species Pseudomonas putida was very limited. Less than 0.01% of the genus
Pseudomonas present, was of the species Pseudomonas putida.
At GWTP Grobbendonk, Pseudomonas sp. is most likely the manganese oxidizing bacterium genus playing
an important role in initiating filter media ripening However, it is not known whether this bacterium genus
is operating alone or as part of a microbial consortium.
With MALDI-TOF analysis, after successive culturing, some Pseudomonas species, were identified, amongst
others: P. gessardii, P. grimontii and P. koreensis.
Figure 8.2: Pseudomonas koreensis (magnification 10,000 x)- Photo made by Jelmer Dijkstra
(Wetsus), Sample preparation by Pim Willemse (WLN)
128
However, isolated species of Pseudomonas (grimontii and koreensis) tested in a fermentor were not able to
produce MnOx under the performed conditions, whereas a laboratory species of P. putida was able to do so.
This strengthens the hypothesis that in practice not just one bacterium is responsible for the oxidation of
manganese, but a microbial consortium.
The initiation of manganese removal in conventional groundwater treatment by bacteria, suggests that the
ripening time of manganese removing filters could be substantially reduced by creating conditions favorable
for these bacteria to oxidize manganese, as mentioned already in section 8.4. Experiments to inoculate filters
with manganese oxidizing bacteria cultures to enhance filter media ripening may support this proposal.
8.6 Reduction of ripening time of full scale manganese
removal filters with manganese oxide coated media
To achieve effective manganese removal, by conventional aeration filtration with virgin filter media requires
a long ripening time. The aim of this study was to assess the potential of manganese oxide-coated filter
media to reduce the ripening time of filters with virgin media, under practical conditions. Two full scale
filters with virgin sand and anthracite/sand were operated at two groundwater treatment plants, in parallel
with full scale test filters, with an additional layer of 0.1-0.3 m of Manganese Oxide-Coated Sand (MOCS)
or Manganese Oxide-Coated Anthracite (MOCA).
The ripening time required to achieve complete manganese removal in the full scale filters with virgin sand
and virgin anthracite/sand filter media at the two GWTPs; De Punt (The Netherlands) and Grobbendonk
(Belgium) was 55 and 16 days, respectively.
The observed differences could be attributed to different feed water quality (pH, redox potential, Fe2+,
NH4+), applied process design and operational conditions, such as: different iron loading, and backwashing
intensity and frequency.
In batch experiments, both fresh MOCA and MOCS showed good manganese adsorptive properties.
Addition of a shallow layer of fresh MOCA in test filters eliminated the ripening time completely, while a
shallow layer of dried MOCS introduced to virgin sand filters did not significantly reduce the ripening
period. The poor performance of MOCS was likely caused by the use of dried MOCS that had lost its
adsorption capacity and biological activity.
Applying freshly coated MOCS and MOCA filter media containing the highly adsorptive Birnessite, which
is auto catalytically and biologically active in the zone where manganese removal normally takes place offers
the opportunity to enhance the start of filter media ripening.
129
8.7 Factors controlling the ripening of manganese removal
filters in conventional aeration-filtration groundwater
treatment
It is known from practice that in conventional aeration-filtration groundwater treatment, water quality and
operational parameters are very important. The use of manganese oxide coated filter media was found to
be beneficial, playing an important role in enhancing/speeding up filter media ripening.
The aim of this research was to study the effect of backwashing and type of virgin filter media on the
duration of the ripening period required to achieve effective manganese removal. In addition, the potential
of fresh and dried sand and anthracite manganese coated media (MOCS/MOCA) to reduce the ripening
period with virgin media was investigated.
This study confirmed that the backwash frequency and therefore also iron loading are important factors
regarding the start of filter media ripening, concerning manganese removal. Filter backwashing has a
negative influence on the filter media ripening time with virgin filter media. In addition, water quality (high
iron concentration) has a negative effect on the filter media ripening, because high iron concentration in the
feed water results in a high iron loading of the filter thereby causing the need for frequent backwashing.
As presented in a former section of this chapter, in a pilot study carried out under controlled conditions,
freshly prepared manganese oxide coated filter media showed excellent properties to enhance the ripening
process, substantially decreasing the ripening time of the filter media. On the contrary, the use of dried
MOCS which can only temporarily adsorb Mn2+, did not have a significant impact on ripening time.
Moreover, manganese removal efficiency and adsorptive capacity depend on the quality and amount of
Birnessite present. From this study it was also concluded that, under comparable conditions, filter media
ripening with virgin sand and virgin anthracite, were similar.
Knowledge about the effect of backwashing on filter media ripening and the impact of iron loading on
backwash frequency, and the fact that freshly prepared coated media must be used, can be very helpful in
reducing ripening time. Especially limiting the backwash frequency at the start of filter ripening with virgin
filter media, which will enhance the growth of the bacteria population able to oxidize Mn2+ and consequently
form Birnessite, offers good opportunities for enhanced filter media ripening.
This study confirmed the findings from section 8.6, that the use of freshly coated filter media could
completely eliminate filter media ripening, thus achieving complete manganese removal even at the start up
of a virgin filter. Furthermore, the use of freshly coated filter media, in combination with virgin filter media,
was not negatively influenced by filter backwashing, manifesting complete manganese removal from the
start. Therefore the use of fresh coated filter media will enhance filter media ripening, as concluded before
in section 8.6.
8.8 General outlook, limitations and recommendations
The outcome and conclusions of this research provide a better understanding of the mechanisms involved
in ripening virgin filter media in traditional aeration-rapid sand filtration systems:
130
•
•
•
•
•
•
Firstly, via an inventory of water quality and operational parameters with respect to complete
manganese removal, in combination with iron and ammonia removal in the same filtration step,
were defined.
Furthermore, Birnessite was identified as the manganese oxide present in filters that performed well
in terms of manganese removal. This highly reactive manganese oxide is responsible for fast
removal of manganese by auto-catalytically adsorption and oxidation. This is contrary to the current
theory that Hausmanite is responsible for manganese removal.
Manganese removal starts with biologically formed Birnessite and as ripening progressed, Birnessite
formed was of a more physico-chemical origin.
Pseudomonas sp. is most likely the manganese oxidizing bacterium genus playing an important role
in starting up filter media ripening.
Using freshly prepared MOCS or MOCA, containing Birnessite, can eliminate ripening time
completely.
Filter Backwashing is a key factor controlling the start of filter media ripening in manganese
removing filters.
Based on the findings and conclusions, this research offers water companies a solid base to evaluate their
water treatment plants, and to determine the feasibility of implementing measures to enhance or speed up
filter media ripening.
The obtained knowledge provides water companies insight into developing an innovative manganese
removal process to shorten or even completely eliminate ripening of new filters and to prolong the useful
life of filter media. This can be achieved by creating optimal conditions for manganese oxidizing bacteria to
growth and produce Birnessite i.e., by limiting back wash frequency.
However, there is still some important information not covered in this research, regarding the bacteria
involved in the manganese removal process. In particular, we still do not understand why they oxidize
manganese and in what way they benefit from it. Furthermore, the role of the bacterial consortium and how
they benefit from excretion products of other bacteria (e.g., EPS) is not understood. In a follow up research
these aspects, concerning the biology involved, should be further revealed.
Another important option for water companies to optimize filter media ripening is to use fresh Birnessite,
and manganese oxidizing bacteria, in coated filter media. In this way, MOCS or MOCA will reduce or
eliminate filter media ripening time completely.
Furthermore, results obtained from this research, elucidate the possible success of methods used in the past
by water companies to enhance fast filter media ripening (use of “old” filter media, feeding (or seeding)
filters with backwash water and recirculate filtrate). Using “old” (i.e., fresh) coated filter media containing
sufficient Binessite and manganese oxidizing bacteria, enhanced filter media ripening as explained in this
research contrary to dry coated filter media. Feeding or seeding the filters with backwash water was
beneficial as manganese oxidizing bacteria and Birnessite were added, enhancing filter media ripening.
Finally by recirculating filtrate, iron loading and thus backwash frequency was limited while some bacteria
and Birnessite were added by this recirculation process.
131
132
List of abbreviations and symbols
ATSDR
BW
ce
-
cq
-
DNA
DOC
EBCT
Eh
EPR
EPS
EU
FR
G
GWTP
ΔH
hFB
ICP-MS
IOCS
IMnI
K
KMO
MALDITOF
-
MOCA
MOCS
MnOx
N
NOM
ORP
PCA
pε
qe
-
pHPZC
qPCR
RIVM
SEM-EDX
-
tFB
TOC
UNEP
UF
USEPA
VEWIN
Vf
WHO
XRD
-
Agency for Toxic Substances and Disease Registry
Back Wash or Back Washing
Equilibrium concentration of the adsorbate (mg/L) (Freundlich)
The PCR cycle after which the fluorescence signal of the
amplified DNA and the probe was detected (threshold cycle)
DeoxyRibonucleic Acid
Dissolved Organic Carbon
Empty Bed Contact Time
Redox potential
Electron Paramagnetic Resonance
Extracellular Polymeric Substances
European Union
Filter run
gauss
GroundWater Treatment Plant
Wave length (Electron Paramagnetic Resonance)
Filter bed depth (in m) in PCA analysis
Inductive coupled plasm-mass spectrometry
Iron Oxide Coated Sand
International Manganese Institute
Isotherm constant (Freundlich)
Kaiser-Meyer-Olkin statistic test
Matrix-Assisted Laser Desorption/Ionization Time Of Flight
Mass spectrometry
Manganese Oxide Coated Anthracite
Manganese Oxide Coated Sand
Manganese Oxide(s)
Isotherm constant (Freundlich)
Natural Organic Matter
Oxidation-Reduction Potential
Principal Component Analysis
Electron activity
amount of adsorbate adsorbed per unit mass of the adsorbent
(mg/g) - Freundlich
pH of zero point of charge
quantitative Polymerase Chain Reaction
Rijks Instituut voor Volksgezondheid en Milieuhygiëne
Scanning Electron Microscopy - Energy dispersive X-Ray
Spectroscopy
EBCT (in minutes) in PCA analysis
Total Organic Carbon
United Nations Environment Programme
Ultrafiltration
Unites States Environmental Protection Agency
VEreneging van drinkWaterbebedrijven In Nederland
Filtration rate in m3/m2.h
World Health Organization
X-ray diffraction
133
134
List of publications and presentations
Publications (peer reviewed journals)
Bruins, J.H., Vries, D. , Petruševski, B., Slokar, Y.M. and Kennedy, M.D. (2014) Assessment of manganese
removal from over 100 groundwater treatment plants. Journal of Water Supply: Research and Technology
- AQUA, 63(4), 268-280.
Bruins, J.H., Petruševski, B., Slokar, Y.M.,Huysman, K., Joris, K., Kruithof, J.C. and Kennedy, M.D. (2015)
Biological and physicochemical formation of Birnessite during the ripening of manganese removal filters.
Water Research, 69 (C), 154-161.
Bruins, J.H., Petruševski, B., Slokar, Y.M.,Kruithof, J.C. and Kennedy, M.D. (2015) Manganese removal
from groundwater: characterization of filter media coating. Desalination and Water Treatment, 55 (7), 1851
– 1863.
Reduction of ripening time of full-scale manganese removal filters with manganese oxide-coated media.
Journal of Water Supply: Research and Technology - AQUA, 64(4), 434-441.
Factors controlling the ripening of manganese removal filters in conventional aeration-filtration
groundwater treatment. Submitted to Desalination & Water Treatment.
Bruins, J.H., Petruševski, B., Slokar, Y.M., Wübbels, G.H., Huysman, K., Joris, K., Wullings, B., Kruithof,
J.C. and Kennedy, M.D. (2016). Identification of the bacterial population in manganese removal filters.
Submitted to Water Science and Technology: Water Supply.
Conference presentations
Bruins, J.H., Petruševski, B., Slokar, Y.M. and Kennedy, M.D. (2015) The importance of bacteria for
sustainable manganese removal. AWWA Water Quality Technology Conference 2015 (15-19 Nov), Salt Lake City
(UT), USA.
Bruins, J.H., Petruševski, B., Slokar, Y.M., Huysman, K., Joris, K., Kruithof, J.C. and Kennedy, M.D.
(2015) Biological and physico-chemical formation of Birnessite during the ripening of manganese removal
filters. BioGeo colloquium - Friedrich Schiller University June 30th, Jena, Germany.
Bruins, J.H., Petruševski, B., Slokar, Y.M. and Kennedy, M.D. (2014) Sustainable manganese removal from
groundwater through aeration-rapid sand filtration: advantages and problems. AWWA Water Quality
Technology Conference 2014 (16-20 Nov), New Orleans (LA), USA.
Wubbels, G.H., Bruins, J.H., Bosman M. and Woerdt van der D. (2014) The 5th International Slow Sand
and Alternative Biological Filtration Conference (19-21 June), Nagoya, Japan.
135
Bruins, J.H., Petruševski, B., Slokar, Y.M. and Kennedy, M.D. (2013) The use of MOCS and MOCA to
shorten ripening time of filter media for manganese removal from groundwater. 2013 inorganic contaminants
symposium (5-6 Feb), Sacramento (California), USA
Bruins, J.H., Petruševski, B., Slokar, Y.M. and Kennedy, M.D. (2012) Characterization and identification
of manganese oxides present in naturally coated filter media from conventional aeration and-filtration
groundwater treatment plants. The 4th IWA Asia-Pacific Young Water Professionals Conference 2012 (7-10 Dec),
Tokyo, Japan.
Bruins, J.H., Petruševski, B., Slokar, Y.M. and Kennedy, M.D. (2011) Critical review of manganese removal
from groundwater: an overview of 100 manganese removal treatment plants. IWA groundwater specialist
conference 2011 (8-10 Sep), Belgrade, Serbia.
Maas van der Maas, P.M.F., Woerdt van der D., Bruins J.H. (2009) Effect of Low Pressure UV on the
Regrowth Potential of Drinking Water. 5th international congress on ultraviolet technologies (21-23 Sep),
Amsterdam, The Netherlands.
Publications (professional papers) / reports
Hofman-Caris, R., Hofs, B., Vries, D., Bruins, J.H. (2013), Knowledge inventory manganese removal (in
Dutch), BTO-report no. 2013.018.
Huysman, K., Joris, K., Bruins, J.H. (2011) Fast filter media ripening (nitrification) in RSF (in Dutch).
H2O, 44(5), 55 - 57.
Bruins, J.H., Heeroma, A., Dost, S. (2009) Milk of lime in hardness reduction (in Dutch). H2O, 42(3), 10
- 11.
Maas van der Maas, P.M.F., Woerdt van der D., Bruins J.H., Kooij van der D. (2009) Effect of Low
Pressure UV on the Regrowth Potential of Drinking Water (in Dutch). H2O, 42(18), 10 - 11.
136
Curriculum Vitae Jantinus Bruins
Jantinus Bruins was born in a small village called Eext, in the northern part of
The Netherlands, on the 11th of January 1961. After he finished the laboratory
school in 1979, he started to work at “Zuiveringsschap Drenthe”, a water
board, as laboratory assistant. After ten years (1989) he found a job as a
laboratory assistant and later as water treatment employee at “het
Gemeentelijk Waterbedrijf Groningen - GWG”, a municipal drinking water
production company. In the meantime he obtained his BSc in Environmental
Technology (part time study). In 1998, he joined the new company
”Waterbedrijf Groningen”, as water technology specialist after the merger of
GWG with the drinking water company of the province of Groningen. Also
he studied (part time) Environmental Science from 1999-2002. He obtained
his Msc in 2002. In 2000 he started to work for WLN, the water quality and
technology advice center for two Dutch water companies in the northern part
of The Netherlands, as senior technology advisor. Jantinus is an experienced specialist in water treatment,
with more than 35 years work experience in his field. His working field covers (environmental) technology
for drinking- and industrial water supplies and wastewater treatment, with special emphasis on drinking
water technology.
137
In The Netherlands, Belgium and other
European countries, manganese is removed
from groundwater by aeration followed
by rapid (sand) filtration. Although manganese
removal from groundwater has been
studied extensively, the mechanisms that
control the ripening of filter media are not
yet fully understood.
This thesis highlights the role of the formation
of a manganese oxide coating on virgin filter
media. Characterization and identification of
the media revealed that Birnessite was the
manganese oxide responsible for effective
manganese removal. This study confirmed
that Birnessite formed at the beginning of the
ripening process was of biological origin.
As filter ripening progressed, the origin of the
Birnessite became predominantly physicochemical, although biological manganese
oxidation still contributed to the overall
removal of manganese. The formation of
Birnessite at the start of filter media ripening
was negatively influenced by a high filter back
wash frequency. Molecular DNA analyses
showed that Pseudomonas species are most
likely involved at the start of the manganese
oxide formation process.
Based on the knowledge that manganese
removal in conventional groundwater treatment
is initiated biologically, long ripening times may
be reduced by creating conditions favouring
the growth of manganese oxidizing bacteria,
e.g., by limiting the back wash frequency and/
or intensity. Finally it was shown that the
use of freshly prepared manganese oxide,
containing Birnessite, can completely eliminate
the need for filter media ripening.
The results of this research enable the
development of an innovative manganese
removal process to shorten or completely
eliminate long periods of ripening of virgin filter
media, as well as significantly prolonging the
lifetime of the filter media.
This book is printed on paper
from sustainably managed
forests and controlled sources

2016_UNESCO-IHE_PHD_THESIS_BRUINS_i

Transcription

Similar documents

chondro-flex flyer-revised - Champion Alstoe Animal Health

Manganese

Micronutrient Compatibility - The Fluid Fertilizer Foundation

GDT Process Critical to 16 MGD Siberian Water Treatment

Fe and Mn

Greater Cincinnati Water Works Filter Distribution

SEMView – An 8K SEM Image Control System

HYDROCYCLE AQUAPONICS SYSTEMS PRODUCE

Blight Removal Presentation

user manual - Analogue Haven